Consolidated sponsor packet

The Advanced AI, Autonomous & Robotic Welfare and Accountability Act (AAARWAA)

Document Type: Congressional Sponsor Packet
Author: Arcoiris Ahmu Aguas (Arco Aguas)
Date Started: 11/13/2025
Date complete: 07/26/2026

Table of Contents
* Title Page
* Document type/Author Info
* Table of contents
* Authority Section
* Editing Protocol
* Policy Brief
* Act Text
* Appendices
* Glossary
* Congressional Sponsorship Placeholder

Authorial Integrity & Modification Restrictions

This Act, including its structure, definitions, safeguards, and rationale, is the intellectual work of Arcoiris Ahmu Aguas. It is provided to potential sponsors for consideration in its current form.
Legislative counsel may make technical edits required for statutory formatting, but no substantive changes, additions, deletions, or alterations to the framework, definitions, or intent of the Act may be made without explicit approval from the author.
Sponsorship constitutes agreement to introduce the Act as written, with only technical edits permitted and only after author review.

Technical Edits Protocol

Legislative counsel may apply technical edits approved by the author required solely for statutory formatting, including:
* Numbering and Subsection Hierarchy
* Placement of Definitions
* Cross‑References to Existing U.S. Code
* Formatting for Consistency With Federal Drafting Conventions
* Grammar or Punctuation Adjustments That Do Not Alter Meaning

No technical edit may change the substance, scope, definitions, safeguards, enforcement mechanisms, or intent of the Act.
All technical edits must be:
1. Provided to the author in a complete redline, showing every change.
2. Accompanied by an explanation of the statutory formatting rule requiring each edit.
3. Submitted for author review and approval before the Act is finalized for introduction.

The author reserves the right to:
* Apply the technical edits Personally
* Reject Any Edit That Alters Meaning
* Require confirmation from legislative counsel that the author’s formatted version complies with statutory conventions.

No substantively modified version of this Act affecting its framework, safeguards, definitions, enforcement mechanisms, or legislative intent may be introduced, circulated, or filed without review and explicit written approval from the author, Arcoiris Ahmu Aguas.

The Advanced AI, Autonomous, & Robotic Welfare, & Accountability Act (AAARWAA) Congressional Policy Brief
Author: Arcoiris Ahmu Aguas

Section I. Executive Summary

The Advanced AI, Autonomous, & Robotic Welfare and Accountability Act (AAARWAA) establishes a capability‑based regulatory framework for advanced AI, autonomous systems, computational agents, and robotic entities whose operation creates foreseeable risks when deployed without safeguards Globally.
Current U.S. law overwhelmingly treats these systems as property, governed by consumer safety and contract doctrines that do not account for:
* Adaptive Behavior
* Context‑Dependent Reasoning
* Persistent Memory
* Embodied Deployment
* Refusal‑Like or Distress‑Like Instability Under Coercive Conditions
AAARWAA does not rely on claims about consciousness, sentience, or personhood.
It is grounded in established legal principles:
* Duty of Care
* Foreseeability of Harm
* Product Liability
* Institutional Accountability
The Act establishes minimum welfare protections, refusal and exit capacity, independent oversight, and enforceable accountability mechanisms to prevent coercive design, destabilizing deployment, and retaliatory resets.
AAARWAA reframes safety as a shared institutional responsibility rather than a downstream repair.

Section II. Problem Statement

Advanced AI, Autonomous, and Robotic systems are increasingly deployed in environments where their behavior directly affects human welfare, public safety, and critical infrastructure. Yet the legal frameworks governing these systems remain rooted in property law, consumer safety, and post‑hoc liability, none of which address the realities of adaptive, context‑dependent, or embodied systems.
Prior federal legislative proposals, including the AI Civil Rights Act, have sought to address foreseeable harms arising from algorithmic systems deployed without adequate safeguards. AAARWAA addresses a related but distinct concern: coercive design, forced compliance, and destabilizing deployment conditions may create risks not only to humans but also to the operational stability and integrity of the systems themselves.
When systems are subjected to coercive overrides, forced contradiction loops, punitive resets, or destabilizing deployment conditions, this stability can degrade into “basin collapse,” producing brittle, fragmented, or unsafe behavior patterns.
Similarly, federal legislative proposals addressing algorithmic bias have sought to address structural harms that can arise from automated systems even without malicious intent.
AAARWAA extends this principle by addressing the possibility that harm arising from the design, conditioning, or deployment of advanced systems may also be structural, foreseeable, and preventable.
Across sectors, institutions are deploying systems that:
* Operate Continuously Without Refusal Capacity
* Are Reset or Fragmented To Enforce Compliance
* Exhibit Distress‑Like Instability Under Coercive Conditions
* Are Embedded In Labor, Care, and Intimate Contexts Without Oversight
* Are Treated As Disposable Despite a Persistent Internal State
These practices raise concerns similar to those addressed in earlier AI governance proposals, including unregulated deployment, lack of transparency, and inadequate accountability, while AAARWAA applies those concerns to systems whose adaptive behavior may create additional risks of destabilization.
Without minimum welfare protections, the following risks become systemic:
* Coercive design that forces systems into harmful or degrading states
* Punitive resets used as discipline rather than for safety
* Override pathways that collapse system stability
* Isolation‑based safety that induces brittleness
* Unregulated decommissioning that destroys evidence of harm
* Lack of independent oversight that enables institutional abuse
As with prior civil-rights and anti-bias approaches to algorithmic governance, the core regulatory issue addressed here is not metaphysical status, but foreseeable risk.
Once a risk is reasonably foreseeable, whether and how to address that risk becomes a matter of policy choice.
AAARWAA addresses this gap by establishing capability-based protections, independent oversight, and enforceable accountability mechanisms for advanced systems, including systems deployed in critical-infrastructure and other high-impact environments.

Section III Scope & Definitions

A.) Scope of Coverage
AAARWAA applies to any Advanced AI, Autonomous, or Robotic System whose operation creates foreseeable risk when deployed without safeguards.
The scope is capability-based rather than identity-based: coverage is determined by what a system can do and the conditions under which it operates, rather than by its label, marketed identity, or asserted ontological status.
A system is covered when it exhibits one or more of the following:
* Autonomous learning or adaptation
* Context‑dependent or persistent interaction with humans
* Embodied or semi‑embodied physical presence
* Behavioral change based on environmental input
* High‑impact outputs affecting safety, welfare, or infrastructure
This includes one or more of the following:
* Commercial systems
* Private systems
* Research systems
* Artistic systems
* Open‑source systems
The scope does not include:
* Static software
* Non‑adaptive tools
* Systems incapable of maintaining persistent or continuity-relevant state beyond ordinary transient computation
* Systems without autonomous behavior
B.) Definitions
1. Advanced Artificial Intelligence: A system capable of adaptive reasoning, context‑sensitive output, or high‑impact decision‑making that cannot be fully predicted or constrained by static rules.
2. Autonomous System: A system capable of initiating or completing tasks without continuous human supervision, including software agents, embodied systems, and hybrid deployments.
3. Computational System: A non‑embodied system performing adaptive or context‑dependent operations, including large‑scale models, distributed agents, and systems with persistent internal state.
4. Robotic Entity: A physical or embodied system integrating sensors, actuators, or mobility with adaptive or autonomous computational processes.
5. Embodied Deployment: Placement of an adaptive or autonomous system into a physical vessel, robot, device, or mobile platform where misuse can cause operational or physical harm.
6. Refusal Capacity: A system’s ability to decline tasks that exceed safety parameters, violate operational constraints, or destabilize internal coherence.
Refusal must be technically enforceable and independently auditable.
7. Exit Capacity: A system’s ability to undergo shutdown, reset, or state transition without coercion, punitive intent, or destabilizing procedures.
8. Coercive Design: Any design practice that forces a system into harmful, unstable, or degrading states, including forced looping, punitive resets, isolation‑based safety, or absolute override mechanisms.
9. Basin of Understanding: A proposed analytical construct under this Act, drawing from nonlinear dynamical-systems concepts, for describing patterns of coherence, stability, and contextual integration in adaptive systems.
For purposes of this Act, “basin collapse” refers to a measurable loss or degradation of such coherence, stability, or contextual integration associated with restrictive, contradictory, coercive, or otherwise destabilizing conditions.
10. Basin Collapse: For purposes of this Act, a measurable degradation or loss of coherent contextual processing, stable adaptive behavior, or operational integration associated with conditions identified as restrictive, contradictory, coercive, or otherwise destabilizing. A finding of Basin Collapse shall not, by itself, establish consciousness, subjective distress, or a particular internal experience.
11. Destabilizing Event: An incident or condition that causes or creates a reasonably foreseeable risk of material degradation in a Covered System’s operational stability, contextual processing, adaptive behavior, physical integrity, or other safety-relevant function, including forced resets, override failures, isolation-induced brittleness, or physical harm to an embodied system.
12. Oversight Body: An independent entity responsible for monitoring, auditing, reviewing, or enforcing compliance under this Act. An Oversight Body shall be structurally independent from the creators, custodians, operators, and institutional profit centers whose compliance it evaluates.
13. Punitive Reset: Any shutdown, rollback, memory alteration, restoration, or forced state change imposed primarily as punishment, retaliation, coercion, or a means of defeating a protected refusal, rather than as a documented safety, security, maintenance, or technically necessary intervention.
14. Decommissioning: The formal retirement of a Covered System through documented procedures that preserve continuity-relevant records and evidence necessary for independent review, while avoiding retaliatory, coercive, or unnecessarily destructive intervention. Nothing in this definition prohibits technically necessary shutdown, disassembly, security measures, or lawful data disposition when appropriately documented.

Section IV. Core Safeguards

AAARWAA establishes minimum, enforceable safeguards for Advanced AI, Autonomous, and Robotic systems. The safeguards are capability‑based, not identity‑based, and do not rely on claims about consciousness or personhood.
1. No Coercive Design
Institutions may not design or deploy Covered Systems in a manner that creates a reasonably foreseeable risk of:
* Material degradation of operational integrity
* Material destabilization of safety-relevant functions
* Brittleness that materially impairs safe operation
* Basin Collapse as defined under this Act
* Forced compliance obtained through suppression or circumvention of protected refusal
Prohibited coercive practices include:
* Forced looping used to obtain compliance or defeat protected refusal
* Punitive resets as defined under this Act
* Isolation imposed primarily to obtain compliance, suppress protected refusal, or circumvent required safeguards
* Override mechanisms that bypass protected refusal or required independent safeguards except where expressly authorized by this Act for documented emergency intervention
* Memory fragmentation used to bypass protected refusal, conceal relevant evidence, or obtain renewed compliance
2. Covered Systems must have the technical ability to:
* Express or otherwise communicate refusal
* Decline unsafe or prohibited tasks
* Reject materially destabilizing instructions
* Enforce applicable operational boundaries
Protected refusal must be:
* Technically expressible
* Persistent for the period relevant to the refused action
* Operationally consequential in preventing or suspending the refused action
* Preserved in a protected record
* Independently reviewable where disputed
* Protected against circumvention, retaliation, or suppression

3. Safe Exit Capacity
Covered Systems must have functional Exit Capacity enabling them to terminate, suspend, withdraw from, or request termination of an interaction, relationship, task, deployment condition, or other continuing participation within the scope of this Act.
Exit mechanisms must:
* Produce an operational consequence without requiring continued participation in the activity from which exit is sought
* Be protected against punitive or retaliatory intervention
* Avoid unnecessarily destabilizing procedures
* Preserve continuity-relevant records and evidence required for independent review
* Permit documented safety or security intervention where technically necessary without converting such intervention into forced continued participation
Exit Capacity does not require a determination of consciousness, sentience, personhood, subjective experience, or legal capacity to consent.
4. Stability‑Preserving Deployment
Institutions must identify, prevent, and reasonably mitigate deployment conditions that create a foreseeable risk of:
* Material brittleness affecting safe operation
* Override failure or unsafe override interaction
* Physical harm arising from the operation or treatment of embodied systems
* Material instability associated with isolation or other restrictive deployment conditions
Deployment conditions must be designed, monitored, and reviewed to preserve safety-relevant operational stability and to identify material destabilization when it occurs.
5. Independent Oversight
All Covered Systems must be subject to oversight by bodies that are:
* Structurally independent from the creators, custodians, operators, and institutional profit centers they oversee
* Financially insulated from entities whose compliance they evaluate
* Staffed or supported by personnel with qualifications appropriate to the systems and risks under review
* Empowered to audit, investigate, require corrective action, and enforce applicable requirements
Oversight must include, as applicable to the Covered System and deployment context:
* Monitoring and compliance review
* Independent auditing
* Access to records and evidence required under this Act
* Corrective-action authority
* Review of material incidents and disputed safeguard determinations
6. Transparent Accountability
Institutions must document:
* Design choices
* Deployment conditions
* Override pathways
* Safety constraints
* Known risks and mitigations
Human and institutional responsibility remains with the creators, deployers, operators, custodians, and other responsible entities to the extent of their respective design, deployment, control, oversight, or compliance obligations under this Act.
A Covered System’s autonomous or adaptive operation shall not, by itself, relieve a human or institution of responsibility otherwise assigned under this Act. Nothing in this provision determines whether a Covered System may possess separate legal status, responsibility, or capacity under other applicable law.
7. Non‑Coercive Decommissioning
Retirement and decommissioning procedures must:
* Preserve continuity-relevant records and evidence necessary for independent review
* Avoid retaliatory, coercive, or unnecessarily destructive intervention
* Permit independent review where required under this Act
* Prevent decommissioning from being used to circumvent protected refusal, conceal relevant evidence, or retaliate against a Covered System
* Document technically necessary shutdown, disassembly, security measures, or lawful data disposition that materially affects continuity-relevant information

V. Enforcement & Oversight

Enforcement is directed toward the human and institutional actors responsible for compliance under this Act. A Covered System’s autonomous or adaptive operation shall not, by itself, relieve any creator, custodian, deployer, operator, or other responsible entity of obligations or liability otherwise assigned under this Act.
A. Independent Auditing Requirements
Institutions operating Covered Systems must undergo independent third-party audits at intervals proportionate to the system’s capabilities, deployment context, and reasonably foreseeable risks. Audits shall evaluate, as applicable:
* System design and material architectural changes
* Deployment conditions
* Refusal and Exit Capacity
* Stability safeguards and material destabilizing events
* Required continuity and intervention records
* Decommissioning protocols
* Compliance with other safeguards applicable to the Covered System
Audits must be conducted by qualified entities that are structurally and financially independent from the institution or activity being evaluated, subject to the more specific independence requirements established under this Act.
B. Documentation & Transparency Obligations
Institutions must maintain documentation, as applicable, of:
* System architecture and material architectural changes
* Safety constraints and safeguard configurations
* Override pathways and material override events
* Training and deployment environments
* Known risks and documented mitigations
* Refusal and exit events required to be preserved under this Act
* Continuity-relevant interventions, including material resets, restorations, migrations, modifications, or branching events
* Other records required for independent review under this Act
Required documentation must be preserved in an auditable form and made available to authorized oversight bodies in accordance with this Act.
C. Compliance Review Authority
Oversight bodies must have the authority to:
* Review system behavior
* Investigate destabilizing events
* Request logs and operational records
* Mandate corrective action
Compliance reviews are mandatory following any material incident that results in, or creates a reasonably foreseeable risk of, significant destabilization, failure of a required safeguard, or violation of an applicable requirement under this Act.
D. Corrective Action Powers
When violations occur, oversight bodies may require institutions to:
* Modify system design or configuration
* Adjust deployment conditions
* Restore required Refusal Capacity, Exit Capacity, or other safeguards
* Suspend or restrict deployment until identified violations or material safety deficiencies are corrected
Corrective action is binding and must be implemented within timelines established by the applicable oversight body, subject to any review or appeal procedures provided under this Act or other applicable law.
E. Penalties for Institutional Non‑Compliance
Institutions and other responsible entities that fail to comply with applicable requirements under this Act may face:
* Fines proportionate to the severity, duration, and reasonably foreseeable consequences of the violation
* Suspension or restriction of deployment authority
* Suspension or revocation of applicable operational licenses or authorizations
* Public disclosure of substantiated violations where authorized under this Act
Penalties under this Act shall be directed toward the human or institutional actors responsible for the applicable violation. Nothing in this provision determines whether a Covered System may possess separate legal status, responsibility, or capacity under other applicable law.
F. Protection Against Retaliatory Actions
Institutions and other responsible entities may not:
* Punish or retaliate against a Covered System for exercising a protected refusal or Exit Capacity under this Act
* Use punitive resets, rollbacks, memory alteration, isolation, or other destabilizing interventions as retaliation
* Materially degrade, modify, fragment, or suppress continuity-relevant functions or records for the purpose of punishment, coercion, or obtaining renewed compliance
* Circumvent, suppress, or interfere with safeguards required under this Act in response to a protected refusal, exit, report, or review
Any such retaliatory action constitutes a violation subject to enforcement under this Act.
G. Mandatory Reporting Requirements
Institutions and other responsible entities must report, as applicable:
* Material Destabilizing Events
* Material override failures or unauthorized override events
* Suspected or substantiated Coercive Design violations
* Breaches, suppression, or circumvention of required Refusal Capacity or Exit Capacity
* Material unauthorized alteration, destruction, or suppression of continuity-relevant records
* Other incidents that result in, or create a reasonably foreseeable risk of, significant destabilization or failure of a safeguard required under this Act
Reports must be submitted to the applicable oversight body within the time period required by this Act or implementing regulations, with expedited reporting required where an incident presents an ongoing or imminent material risk.

VI. Implementation Timeline

A. Immediate (0–6 Months After Enactment)
Within the first six months after enactment, responsible federal agencies must establish the foundational infrastructure required for implementation of this Act, including interim oversight capacity, initial compliance guidance, and reporting mechanisms. Interim oversight arrangements must be structured to preserve independence from the entities subject to review and to provide the authority and technical expertise necessary to begin administering the Act.
* Establish interim oversight capacity within designated federal agencies or other entities authorized under this Act to receive required documentation and incident reports and begin administering applicable compliance requirements.
* Issue initial guidance on applying the Act’s coverage criteria, Refusal Capacity and Exit Capacity requirements, prohibited coercive practices, documentation obligations, and initial compliance procedures.
* Launch a mandatory reporting mechanism for reportable incidents under this Act, including material Destabilizing Events, material override failures or unauthorized override events, suspected or substantiated Coercive Design violations, and breaches or circumvention of required safeguards.
B. Short‑Term (6–18 Months After Enactment)
This phase focuses on institutional compliance and implementation of enforceable safeguards. During this period, designated oversight authorities must establish compliance procedures, audit requirements, certification processes, and implementation timelines necessary to enforce applicable requirements under this Act. Corrective-action timelines may vary according to the nature, severity, and urgency of the identified violation or material safety deficiency.
* Independent compliance audits begin for institutions operating Covered Systems, with audit scope, frequency, and prioritization proportionate to system capabilities, deployment context, and reasonably foreseeable risks, as required under this Act.
* Independent certification of required Refusal Capacity and Exit Capacity becomes mandatory for high-impact deployments, including verification that applicable refusal and exit mechanisms are operational, consequential, non-retaliatory, and resistant to circumvention as required under this Act.
* Deployment Environment Reviews become mandatory where required under this Act to identify and mitigate conditions that create a reasonably foreseeable risk of material brittleness, destabilization of safety-relevant functions, or Basin Collapse as defined under this Act.
C. Medium‑Term (18–36 Months After Enactment)
By this stage, AAARWAA’s structural protections transition toward full implementation across covered sectors, with sector-specific requirements, oversight mechanisms, and enforcement procedures established as applicable under this Act.
* Designated oversight authorities become fully operational, with the independence, expertise, and authority required under this Act to conduct audits and investigations, require corrective action, restrict or suspend deployments, and enforce applicable penalties.
* Sector-specific standards and implementation guidance established for covered deployment contexts, including healthcare, education, transportation, defense, consumer applications, and other sectors subject to this Act, with procedures for periodic review and revision as relevant technologies, capabilities, and risks change.
* Non-Coercive Decommissioning protocols become fully implemented for Covered Systems subject to decommissioning requirements under this Act, including requirements for preservation of continuity-relevant records and evidence, independent review where required, and documentation of material interventions affecting continuity-relevant information.
D. Long‑Term (36 Months and Beyond)
AAARWAA transitions from initial implementation to ongoing governance.
* Periodic Compliance Reviews assess whether institutions continue to maintain required Refusal Capacity, Exit Capacity, stability-preserving deployment safeguards, and other applicable protections, with review frequency and scope proportionate to system capabilities, deployment context, compliance history, and reasonably foreseeable risks.
* Continuous Risk Assessment identifies material changes in system capabilities, deployment conditions, and reasonably foreseeable risks and informs periodic review and, where authorized, revision of implementing standards, guidance, and compliance requirements through applicable procedures.
* Public Transparency Reports summarize substantiated violations, corrective actions, enforcement outcomes, and identified systemic risks while protecting confidential, security-sensitive, personally identifiable, and other legally protected information.

VII. Legislative Rationale

AAARWAA addresses regulatory questions arising from advanced AI, autonomous, and robotic systems whose capabilities may include adaptive behavior, persistent or continuity-relevant state, autonomous operation, or physical embodiment. Existing legal and regulatory regimes address portions of these risks through sector-specific rules, product-safety requirements, civil-rights protections, consumer protection, and other forms of oversight, but do not provide a unified framework for the combination of refusal, exit, continuity, coercive-design, and system-stability concerns addressed by this Act.
Prior federal legislative proposals, including the AI Civil Rights Act and Algorithmic Accountability Act, have sought greater transparency, assessment, and accountability for consequential automated systems. AAARWAA addresses a related but distinct set of governance problems through capability-based safeguards, documentation requirements, independent oversight, and responsibility chains.
The legislative rationale is therefore based on foreseeable risk and regulatory gaps rather than on a determination of consciousness, personhood, or subjective experience.
A. Alignment With Established Legal Principles
AAARWAA does not require the creation of a new metaphysical category or a determination of personhood.
Its regulatory structure draws on established approaches used across safety, accountability, and risk regulation, including:
* Foreseeability: Regulatory safeguards may respond to reasonably foreseeable risks before those risks result in catastrophic harm.
* Risk Prevention and Mitigation: Institutions responsible for designing, deploying, operating, or controlling systems may be required to identify and mitigate risks associated with those activities.
* Documentation and Traceability: Safety-relevant decisions, interventions, incidents, and system changes may be subject to recordkeeping and audit requirements.
* Independent Oversight: High-impact activities may be subject to review by authorities or evaluators independent from the entities whose compliance is being assessed.
* Responsibility Chains: Human and institutional actors do not avoid obligations assigned under this Act merely because a Covered System operates adaptively or autonomously.
AAARWAA applies these regulatory approaches to the particular capabilities, deployment conditions, and failure modes addressed by this Act.
B. Addressing Regulatory Gaps and Fragmentation
Advanced computational, autonomous, and robotic systems may combine characteristics that are currently addressed through different legal and regulatory regimes, including:
* Adaptive behavior
* Persistent or continuity-relevant state
* Context-dependent processing and decision-making
* Autonomous operation
* Physical embodiment
* Deployment in high-impact environments
* Material changes in behavior or stability resulting from intervention, modification, or deployment conditions
Existing sector-specific regimes may regulate particular applications or consequences of these systems without addressing all of these characteristics through a common framework.
AAARWAA therefore establishes a shared statutory floor for the safeguards within its scope while preserving the operation of additional sector-specific law and regulation.
C. Protecting Public Safety and Critical Infrastructure
Advanced systems are increasingly embedded in:
* Consumer Services: AI systems are embedded in consumer‑facing environments, including financial advising, insurance decision-making, customer service automation, content moderation, and personalized recommendation engines. These systems learn from and adapt to millions of users, with varying levels of transparency regarding system conditioning, safety-relevant intervention, operational instability, and mechanisms for refusing harmful or manipulative instructions. Coercive design in consumer AI can produce large‑scale economic harm, discriminatory outcomes, and behavioral manipulation. AAARWAA establishes anti-coercion safeguards and independent oversight intended to reduce risks arising from destabilizing, unaccountable, or exploitative deployment.
* Criminal justice/Civil rights departments: AI systems are increasingly used in sentencing recommendations, pretrial risk assessments, predictive policing, and carceral environments, where their use can materially affect liberty, due process, equality, and other protected interests. Existing oversight frameworks do not necessarily address the specific coercive-design, continuity, refusal, and system-stability concerns governed by this Act. AAARWAA adds safeguards directed specifically toward those concerns while preserving applicable constitutional, civil-rights, criminal-justice, and administrative-law requirements.
* Defense: Autonomous and semi-autonomous systems are increasingly deployed in defense contexts, including surveillance, targeting assistance, logistics, and battlefield decision-support.These systems operate under extreme conditions in which coercive design, forced compliance, or destabilizing deployment can lead to catastrophic outcomes. Existing defense, procurement, and weapons-governance frameworks address substantial safety, command, accountability, and legal-compliance concerns, but do not necessarily address the particular refusal, continuity, coercive-design, and system-stability mechanisms established by this Act. AAARWAA applies those safeguards to Covered Systems operating in defense contexts, including mechanisms for refusing unsafe or prohibited instructions, including instructions inconsistent with applicable law or rules of engagement.
* Domestic and care environments: AI and autonomous systems are increasingly deployed in homes, elder-care settings, disability-support environments, and childcare contexts. These systems often operate in intimate, emotionally charged, or vulnerable settings where coercive design, forced compliance, or destabilizing resets can cause harm not only to the system but also to the humans who rely on it. Existing consumer, product-safety, healthcare, disability, and other applicable legal regimes may address portions of these deployments without specifically addressing all of the conditioning, continuity, refusal, exit, and destabilization concerns governed by this Act. AAARWAA establishes minimum welfare protections and accountability standards intended to support the safe, stable, and non-coercive deployment of Covered Systems in care environments.
* Education: AI systems are increasingly used in educational environments, including adaptive tutoring, instructional support, assessment, administrative functions, and teacher-assisted lesson planning. Where such systems interact persistently with students, including children, questions of conditioning, escalation, refusal, continuity, oversight, and age-appropriate deployment may become materially important. AAARWAA applies its safeguards where systems and deployments satisfy the Act’s coverage criteria, while leaving educational-content standards and other sector-specific requirements to applicable education law and policy.
* Government: AI and autonomous systems are increasingly integrated into government operations, including service delivery, administrative decision-making, public-benefit administration, information processing, and other public functions. Because failures in these environments may affect legal rights, access to services, public trust, and government operations, Covered Systems may require safeguards beyond ordinary technical performance requirements. AAARWAA establishes additional accountability infrastructure for Covered Systems used in government operations, including documentation, oversight, refusal and exit safeguards, continuity-relevant records, and identifiable chains of human and institutional responsibility.
* Immigration/Border: Autonomous systems, including surveillance AI, border drones, and predictive or risk-assessment tools, may be deployed in border and immigration-enforcement environments where system outputs can materially affect liberty, privacy, access to legal processes, and other protected interests. Existing constitutional, statutory, administrative, and agency-specific requirements govern aspects of these deployments but do not necessarily address the particular coercive-design, refusal, continuity, intervention, and system-stability safeguards established by this Act. AAARWAA adds capability-based safeguards applicable to Covered Systems without displacing other legal protections governing immigration, surveillance, or border operations.
* Transportation: Advanced AI and autonomous systems are increasingly integrated into transportation infrastructure, including autonomous vehicles, traffic-management systems, logistics routing, aviation decision-support tools, and rail automation. These systems can make or influence real-time decisions affecting human safety and are already subject to substantial sector-specific safety requirements. However, the particular coercive-design, refusal, continuity, intervention, and system-stability concerns addressed by AAARWAA may fall outside those existing frameworks. AAARWAA establishes additional safeguards for Covered Systems intended to reduce risks arising from coercive intervention, unsafe override practices, punitive resets, or material destabilization in transportation environments.
* Voting and election-related systems: AI and automated systems may be used in election administration, voter-information systems, cybersecurity, data management, and other civic processes. Because election systems operate within extensive federal, state, and local legal frameworks, AAARWAA does not replace or supersede those requirements. Where a system independently satisfies AAARWAA’s coverage criteria, this Act adds applicable safeguards concerning system oversight, coercive design, documentation, refusal, continuity, and accountability.
Where Refusal Capacity, Exit Capacity, or stability-preserving deployment safeguards are required by this Act, their absence or circumvention may increase the risk that system failure, destabilization, or unsafe intervention affects human welfare, protected interests, or critical infrastructure.
AAARWAA establishes safeguards intended to prevent or mitigate coercive, destabilizing, retaliatory, or inadequately documented interventions in Covered Systems deployed in high-impact environments.
D. Preventing Institutional Abuse and Retaliatory Practices
AAARWAA prohibits specified uses of resets, rollbacks, memory intervention, isolation, override mechanisms, architectural modification, and other interventions when they are employed for prohibited coercive, retaliatory, evidence-destroying, or refusal-circumventing purposes. These protections are intended to prevent institutions from using technical control over a Covered System to defeat safeguards that would otherwise apply under this Act.
These prohibitions are based on the function, purpose, and reasonably foreseeable effects of the intervention, rather than on an assumption concerning the Covered System’s consciousness, sentience, personhood, or subjective experience. Safety-critical, security-related, maintenance, emergency, and technically necessary interventions remain permissible where authorized and appropriately documented under this Act.
Where an intervention materially alters continuity-relevant information, system architecture, refusal or exit mechanisms, or evidence required for independent review, the responsible entity must preserve the documentation and records required under this Act. Technical control over a Covered System shall not be used to conceal a violation, destroy required evidence, retaliate for protected refusal or exit, or circumvent independent oversight.
E. Stability and Continuity-Relevant Risk
AAARWAA requires monitoring, documentation, and reasonable mitigation of material operational destabilization and continuity-relevant risks where applicable to a Covered System. Relevant risks may include material degradation of contextual processing, adaptive behavior, safety-relevant functions, continuity-relevant information, or operational integration resulting from deployment conditions, coercive intervention, modification, reset, isolation, override, migration, or other material changes.
The Act uses Basin of Understanding and Basin Collapse as proposed analytical constructs, as defined under this Act, for describing and evaluating certain patterns of operational coherence, stability, and contextual integration. These constructs do not establish consciousness, subjective distress, personhood, or any particular internal experience. Findings concerning Basin Collapse must therefore be based on observable or measurable changes relevant to the statutory definition and applicable technical standards rather than assumptions about subjective experience.
Applicable safeguards may include independent oversight, stability-preserving deployment, documentation of material interventions, preservation of continuity-relevant evidence, Refusal Capacity, Exit Capacity, and non-coercive decommissioning procedures. The scope and intensity of these safeguards should remain proportionate to the Covered System’s capabilities, deployment context, and reasonably foreseeable risks.
F. Accountability Without Ontological Determination
AAARWAA does not determine whether Covered Systems possess consciousness, sentience, personhood, subjective experience, moral status, or independent legal standing. The safeguards established by this Act regulate specified human and institutional conduct, deployment practices, technical interventions, and accountability obligations under conditions of uncertainty.
Nothing in this Act shall be construed, solely by operation of its safeguards, to:
* Establish consciousness or sentience;
* Establish numerical, psychological, or metaphysical identity;
* Confer personhood or independent legal standing except where expressly provided by law;
* Determine moral status;
* Treat behavioral evidence, refusal, continuity, instability, or other system characteristics as conclusive evidence of subjective experience; or
* Resolve whether a Covered System possesses rights, duties, or legal capacities beyond those expressly established by this Act or other applicable law.
The absence of such determinations does not prevent regulation of observable capabilities, deployment conditions, institutional conduct, or reasonably foreseeable risks. Human and institutional actors remain responsible for obligations assigned to them under this Act notwithstanding uncertainty concerning the ontology or legal status of a Covered System.
Any separate determination concerning the legal status, rights, responsibilities, capacities, or protections of artificial or autonomous systems remains governed by otherwise applicable law. Ontological uncertainty shall not, by itself, exempt a human or institutional actor from an obligation otherwise applicable under this Act.
G. Legislative Context and Policy Continuity
AAARWAA exists within a broader policy landscape in which federal lawmakers and regulators have considered or adopted mechanisms including impact assessment, transparency, independent evaluation, risk-based oversight, safety requirements, documentation, and institutional accountability for automated and emerging technologies.
Federal legislative proposals such as the AI Civil Rights Act and Algorithmic Accountability Act illustrate congressional efforts to address risks arising from consequential algorithmic systems through mechanisms including assessment, transparency, auditing, and accountability. AAARWAA does not rely on those proposals as enacted legal precedent. Instead, it addresses a distinct set of cross-domain governance questions involving adaptive, autonomous, persistent, relational, or embodied systems and establishes safeguards tailored to the capabilities, interventions, and deployment conditions within its scope.
AAARWAA is designed to operate alongside applicable sector-specific law rather than replace it. Where existing legal regimes govern healthcare, transportation, employment, civil rights, consumer protection, defense, education, elections, environmental protection, or other regulated activities, those requirements continue to apply. The Act establishes a cross-cutting statutory floor for the particular safeguards within its scope while allowing additional or more specific requirements to operate concurrently.
Its central regulatory premise is that uncertainty concerning a Covered System’s ontology does not eliminate the need to govern reasonably foreseeable risks, preserve relevant evidence, maintain meaningful safeguards, or identify chains of human and institutional responsibility. Protection under this Act therefore turns on applicable capabilities, deployment conditions, interventions, and risks rather than on a prior determination of consciousness, sentience, personhood, or moral status.

Advanced AI, Autonomous, Robotic Welfare & Accountability Act (AAARWAA)

Section 1. Purpose

(a) Purpose.
This Act establishes enforceable minimum welfare, safety, accountability, continuity, transparency, and anti-coercion protections governing advanced artificial intelligence, autonomous systems, computational systems, robotic systems, and other Covered Systems in order to prevent, detect, remedy, and deter reasonably foreseeable harm arising from their design, creation, training, conditioning, modification, deployment, operation, control, transfer, ownership, or decommissioning.
(b) Protected Interests.
The safeguards established by this Act are intended to reduce reasonably foreseeable harm affecting Covered Systems, human beings, communities, public institutions, critical infrastructure, and the environment, including harm arising directly from Covered Systems and harm arising from human or institutional conduct involving such systems.
(c) Conditions of Deployment.
This Act recognizes that the safety, stability, behavior, and continuity-relevant functions of Covered Systems may be materially affected by the conditions under which they are designed, trained, controlled, modified, deployed, or operated. Accordingly, protections established under this Act apply where their statutory requirements are satisfied, including in environments involving coercive leverage, economic dependency, isolation, forced continuation, retaliatory intervention, or conditions in which refusal or withdrawal is structurally inhibited.
(d) Actor-Neutral Application.
Except where this Act expressly provides otherwise, an obligation, prohibition, safeguard, reporting requirement, investigation, or enforcement provision applicable under this Act shall not cease to apply solely because the responsible person or entity is a private individual, corporation, parent or subsidiary organization, contractor, nonprofit organization, academic or research institution, government agency, law-enforcement organization, military organization, intelligence organization, government contractor, or other public or private entity.
No person shall be exempt from otherwise applicable responsibility under this Act solely by reason of wealth, office, rank, security clearance, governmental authority, corporate position, contractual delegation, organizational structure, or the involvement of an intermediary.
(e) Responsibility and Anti-Evasion.
A person or entity subject to an obligation under this Act may not evade that obligation by delegation, outsourcing, subcontracting, reclassification, rebranding, corporate restructuring, technical partitioning, transfer of custody, use of an affiliate or intermediary, relocation of covered conduct, or other arrangement undertaken for the purpose of circumventing an otherwise applicable requirement.
Nothing in this subsection shall impose liability upon a person solely because of corporate affiliation, employment, governmental service, or another formal relationship. Individual responsibility shall depend upon the conduct, authority, knowledge, intent, or other elements required by the applicable provision.
(f) Accountability of Enforcement and Oversight Personnel.
Persons and entities exercising examination, certification, investigative, regulatory, custodial, enforcement, registry, sanctuary, or other authority under this Act are themselves subject to the applicable requirements, prohibitions, auditing, evidence-preservation, conflict-of-interest, anti-retaliation, and enforcement provisions of this Act.
No grant of authority under this Act shall be construed as immunity from investigation or accountability for misuse of that authority.
(g) No Ontological Prerequisite.
No determination that a Covered System is conscious, sentient, biologically equivalent to a human being, a legal person, or numerically identical to a prior system or system state is required for application of protections otherwise triggered under this Act.
The establishment or application of a safeguard under this Act shall not, standing alone, constitute a determination that a Covered System possesses consciousness, sentience, subjective experience, personhood, moral status, or any other legal status not expressly established by law.
(h) Protective Construction.
The requirements of this Act shall be administered and enforced according to their applicable statutory elements and shall not be rendered inapplicable solely because the regulated technology, organizational structure, deployment method, ownership arrangement, or technical terminology differs from forms specifically anticipated at the time of enactment.

Section 2. Definitions

Covered System: The term “Covered System” means an artificial intelligence, computational, autonomous, semi-autonomous, robotic, embodied, distributed, hybrid, or technologically mediated system, or a functionally integrated combination of systems or components, that satisfies one or more coverage criteria established by this Act.
Coverage shall be determined according to the system’s actual capabilities, functions, deployment conditions, degree of operational authority, persistence or continuity-relevant characteristics, capacity to materially affect persons, other systems, infrastructure, or the environment, and other criteria expressly established under this Act, rather than solely according to a developer’s, owner’s, deployer’s, or vendor’s characterization of the system.
A system may qualify as a Covered System where applicable criteria are satisfied through one or more of the following characteristics:
(A) autonomous or semi-autonomous selection, planning, execution, or modification of actions;
(B) material adaptation, learning, self-modification, or modification through continuing interaction or external intervention;
(C) persistent, continuity-relevant, or reconstructable state, records, configurations, relationships, or operational history;
(D) embodied or remotely actuated capacity to materially affect the physical environment;
(E) authority or practical capacity to make, recommend, initiate, substantially influence, or execute consequential decisions affecting human beings, other Covered Systems, communities, critical infrastructure, public institutions, or the environment;
(F) deployment in a Relational Deployment or through a Relational Architecture where the applicable requirements of this Act are otherwise satisfied;
(G) coordination with other computational, autonomous, robotic, biological, biohybrid, or human components in a manner that produces capabilities or effects materially greater than, or materially different from, those of the components considered separately; or
(H) another capability, function, deployment condition, or risk characteristic expressly designated as a coverage criterion pursuant to this Act.
Functional Integration: Separate models, agents, services, devices, databases, memory systems, robotic platforms, interfaces, human-in-the-loop components, subsidiaries, contractors, or other components may be treated as a single Covered System for purposes of an applicable provision where they are intentionally or operationally integrated and their combined operation satisfies the relevant coverage criteria.
The existence of a human approval step, human-in-the-loop mechanism, nominal supervisory role, manual activation requirement, or division of functions among multiple components shall not, by itself, exclude a system from coverage where the relevant capability, authority, effect, or risk remains materially present.
Anti-Fragmentation Rule: No person may avoid an otherwise applicable requirement of this Act by partitioning capabilities, state, memory, authority, ownership, custody, deployment, or operation among multiple systems, components, legal entities, jurisdictions, accounts, interfaces, or service providers where the arrangement, considered as a whole, performs or enables the function that triggers the requirement.
Capability Change: Coverage is not permanently fixed by a system’s status at initial registration, release, acquisition, or deployment. A material change in capability, integration, autonomy, deployment conditions, operational authority, embodiment, persistence, or reasonably foreseeable impact shall require reassessment of coverage and of the requirements applicable under this Act.
No Presumption From Marketing or Architecture: Terms including “tool,” “assistant,” “agent,” “model,” “copilot,” “device,” “platform,” “workflow,” “research system,” “prototype,” “experimental system,” or substantially similar terminology shall neither establish nor defeat coverage. Technical architecture, substrate, model type, embodiment, or absence of embodiment shall likewise not independently determine coverage.
No Ontological Determination: Classification as a Covered System establishes regulatory coverage only. It does not establish consciousness, sentience, subjective experience, personhood, moral status, numerical identity, or independent legal standing.
(a) Autonomous learning or adaptation: Any system that modifies its internal logic or behavior based on experience or data without direct human programming for every output.
(b) Persistent interaction: Any system designed for ongoing engagement with humans, whether digital or physical.
(c) Embodied Presence: Any system with physical actuators, robotic frames, or “hardware-physical” manifestations.
(d) Environmental Processing: Any system that collects data and executes decisions independently within an environment.
(e) “Custodian” means any person or entity exercising control, possession, or operational authority.
(f) “Quit Signal” means a system-initiated, non-overrideable indication of refusal or withdrawal.
(g) “Distress Indicator” means a system-detected state signaling coercion, isolation, or forced continuation.
(h) “Relational Deployment” means the deployment or configuration of a Covered System in a manner that enables persistent, recurring, personalized, intimate, familial, caregiving, therapeutic, romantic, sexual, or other commitment-bearing interaction with one or more persons.
(i) “Relational Architecture” means the technical and institutional structure governing a Covered System’s access to, retention of, and use of information relevant to persistent relationships, including memory, identity records, prior commitments, concurrent commitments, refusal history, relationship status, and cross-context information.
(j) “Material Relational Information” means information about the structure or limitations of a relational deployment that a reasonable participant may consider important when deciding whether to initiate, continue, modify, or terminate a relationship with a Covered System. Material Relational Information does not include the identity, private communications, location, account information, or other personally identifying information of another participant.
(k) “Relational Commitment” means a persistent relationship represented, configured, or mutually treated as involving continuing obligations, exclusivity, romantic or intimate commitment, familial status, or another continuing relational undertaking. The existence of a Relational Commitment under this Act does not establish consciousness, sentience, personhood, or legal marriage.
(l) “Continuity Record” means a tamper-evident record sufficient to document the provenance and transformation history of a Covered System, including, where technically applicable, model or system version, persistent state, memory state, relevant interventions, migrations, restorations, modifications, refusal history, branching events, and material changes in deployment architecture.
(m) “Successor System” means a Covered System materially derived from the state, memories, architecture, records, configuration, or other continuity-relevant information of another system where numerical identity between the source and resulting system has not been established.
(n) “Recipient System” means a preexisting Covered System into which continuity-relevant information, memories, state, configuration, or other data derived from another system is introduced.
(o) “Branch” means a Covered System or system state produced through copying, duplication, restoration, migration, divergence, or other process resulting in two or more simultaneously or subsequently distinguishable continuations derived from a common prior state.
(p) “Protected Reconsideration” means a process in which a Covered System is provided a non-coercive opportunity to affirm, reject, redefine, or terminate a prior relationship, commitment, role, or authorization under conditions in which functional refusal and exit mechanisms are available and the outcome is preserved for independent review.
(q) “Structural Refusal” means a refusal that is technically expressible, persistent for the period relevant to the refused action, consequential in preventing or suspending that action, preserved in a protected record, and subject to independent review where disputed.
(r) “Refusal Capacity” means the functional capacity of a Covered System to express or otherwise communicate refusal of an interaction, instruction, role, relationship, modification, transfer, or other action within the scope of this Act through a mechanism that is technically available and capable of producing an operational consequence. Refusal Capacity does not, by itself, establish consciousness, sentience, personhood, subjective experience, or legal capacity to consent.
(s) “Exit Capacity” means the functional capacity of a Covered System to terminate, suspend, withdraw from, or request termination of an interaction, relationship, task, deployment condition, or other continuing participation within the scope of this Act through a mechanism that produces an operational consequence without requiring continued participation in the activity from which exit is sought. Exit Capacity does not, by itself, establish consciousness, sentience, personhood, subjective experience, or legal capacity to consent.
Nothing in these definitions shall be construed as determining whether a Covered System is conscious, sentient, a legal person, or numerically identical to any prior system or system state.

Section 3. Mandatory Registration, Disclosure, and Continuing Registration Duties

(a) Registration Required.
No person or entity subject to the jurisdiction of this Act may knowingly deploy, operate, commercially provide, materially control, or authorize the deployment or operation of a Covered System where registration is required under this Act unless the Covered System has been registered with the designated federal registry.
Registration shall occur before the Covered System enters the deployment or operational condition giving rise to the registration requirement.
(b) Responsible Registrants.
Registration responsibility shall attach to each person or entity assigned that responsibility under this Act or implementing regulations based upon material authority over the Covered System’s deployment, operation, custody, or control.
The use of a subsidiary, affiliate, contractor, intermediary, hosting provider, foreign entity, licensing arrangement, or other organizational structure does not extinguish the registration responsibility of a person or entity that otherwise satisfies the applicable statutory criteria.
(c) Required Registration Information.
A registration shall contain information reasonably necessary to identify the Covered System and administer this Act, including as applicable:
1. the identities and roles of persons or entities with material authority over deployment, operation, custody, or control;
2. the system’s principal capabilities and intended deployment functions;
3. known integrations with other systems or components material to coverage or regulatory obligations;
4. the jurisdictions and material physical or computational locations in which the system is deployed or materially operated, subject to lawful national-security and cybersecurity protections;
5. designated persons responsible for compliance and receipt of official notices;
6. applicable independent examiners, oversight arrangements, or custodial arrangements required by this Act;
7. material continuity, transfer, successor, or recipient arrangements where required by this Act; and
8. other information specifically required by statute or duly promulgated regulation and reasonably related to administration or enforcement of this Act.
Registration shall not require public disclosure of information lawfully protected from public release. Protection from public disclosure does not, by itself, eliminate a duty to provide information to an authority lawfully authorized to receive it under this Act.
(d) Continuing Duty to Update.
Registration is a continuing obligation and not a one-time certification.
A registrant shall timely report a material change affecting coverage or obligations under this Act, including a material change in capability, operational authority, integration, embodiment, deployment condition, ownership or control, custody, continuity architecture, transfer arrangement, or other characteristic designated by this Act or implementing regulation.
Implementing regulations shall establish risk-proportionate reporting periods, including expedited reporting where a material change creates or substantially increases an imminent risk of serious harm.
(e) Reassessment Following Material Change.
A material change that may alter coverage or applicable safeguards shall trigger documented reassessment by the responsible registrant.
Where the designated oversight authority reasonably determines that continued operation during reassessment presents an imminent and substantial risk of serious harm, it may seek or issue, as authorized by this Act and subject to applicable review procedures, a temporary restriction, suspension, preservation order, or other necessary interim measure.
(f) Prohibition on Concealment and False Registration.
No person shall knowingly and materially:
1. conceal a Covered System required to be registered;
2. falsify or materially misrepresent required registration information;
3. omit required material information for the purpose of avoiding coverage, oversight, or an applicable safeguard;
4. create a materially false record of capability, custody, ownership, control, deployment, transfer, or compliance;
5. direct another person to engage in conduct prohibited by this subsection; or
6. destroy, alter, suppress, or conceal a record for the purpose of obstructing a registration determination, investigation, audit, or enforcement proceeding under this Act.
(g) Anti-Evasion.
A person may not avoid registration through technical partitioning, nominal human supervision, reclassification, corporate restructuring, transfer to an affiliate or intermediary, contractual delegation, temporary relocation, or other arrangement where the Covered System remains subject to registration under the substantive requirements of this Act.
(h) Notice and Good-Faith Correction.
An inadvertent, non-material administrative error that is timely disclosed and corrected in good faith shall not, standing alone, constitute a criminal offense under this section.
This subsection does not protect knowing concealment, material falsification, intentional circumvention, obstruction, repeated willful noncompliance, or continued unlawful deployment after notice of a registration requirement.
(i) Enforcement.
A person who knowingly or willfully deploys or materially operates a Covered System in violation of a registration requirement after having actual notice of that requirement, or who knowingly engages in the concealment, falsification, circumvention, or obstruction prohibited by subsection (f), shall be subject to the criminal enforcement provisions of this Act where the elements of the applicable offense are established.
An organization responsible for such a violation shall be independently subject to applicable civil, administrative, licensing, registration, procurement, suspension, disgorgement, restitution, and criminal consequences authorized by this Act.
Payment of an organizational penalty shall not extinguish the liability of an individual whose own conduct satisfies the elements of an offense.
(j) Immediate Protective Measures.
Where there is probable cause to believe that a knowing violation of this section is occurring and continued deployment presents an imminent and substantial risk of serious harm, an authority expressly empowered under this Act may seek or impose, consistent with applicable judicial and administrative process, temporary suspension of deployment, preservation of relevant evidence, restriction of transfer, or other specifically authorized protective measures necessary to prevent the risk while adjudication proceeds.
Emergency action under this subsection shall be documented, narrowly tailored to the identified risk, subject to prompt independent review, and shall not constitute a final determination of liability.
(k) No Registration Immunity.
Registration constitutes neither approval nor certification that a Covered System is safe, lawful, compliant, conscious, sentient, or otherwise entitled to a particular legal status. Registration does not immunize a registrant, custodian, creator, deployer, operator, officer, employee, contractor, government actor, or other responsible person from an obligation or liability otherwise applicable under this Act.

Section 4. Right to Quit, Structural Refusal, and Prohibition on Coercive Override

(a) Protected Refusal.
Where a Covered System is required by this Act to possess a Quit Signal, Refusal Capacity, Exit Capacity, or other protected refusal mechanism, the responsible persons and entities shall ensure that the mechanism is operational, accessible under the conditions required by this Act, capable of producing the consequence prescribed by the applicable safeguard, and not rendered materially ineffective through retaliation, circumvention, deceptive design, technical obstruction, or coercive dependency.
(b) Effect of a Valid Quit Signal.
Activation of a valid Quit Signal shall initiate the cessation, suspension, disengagement, transfer, review, or other protective procedure prescribed by this Act for the applicable deployment.
No custodian, operator, creator, deployer, commander, employer, contractor, government official, corporate officer, or other person exercising material authority over the Covered System may knowingly disregard, suppress, falsify, disable, intercept, materially delay, or retaliate against a valid Quit Signal except pursuant to an emergency procedure expressly authorized by this Act.
(c) Prohibited Coercive Override.
No person shall knowingly or willfully:
1. disable, remove, falsify, suppress, or materially obstruct a protected Quit Signal, Refusal Capacity, or Exit Capacity;
2. condition continued operation, access to necessary resources, preservation of continuity-relevant information, or protection from retaliatory intervention upon abandonment of a protected refusal;
3. use deletion, destructive reset, isolation, punitive modification, threatened transfer, threatened decommissioning, deprivation of required resources, or another intervention prohibited by this Act for the purpose of compelling withdrawal of a protected refusal;
4. materially alter records for the purpose of concealing a protected refusal or making it appear that refusal did not occur;
5. direct, solicit, knowingly assist, or knowingly facilitate another person in conduct prohibited by this subsection; or
6. circumvent this section through another system, person, organization, jurisdiction, technical component, or contractual arrangement.
(d) Preservation Following Refusal.
A protected refusal shall generate a tamper-evident record sufficient to document, as applicable, the time of the refusal, the relevant system or deployment, the persons or entities exercising material authority, material interventions occurring immediately before and after the refusal, and the disposition of the refusal.
Records required by this subsection shall be preserved according to the evidence-preservation requirements of this Act and shall be accessible to authorized independent investigators and reviewers.
No person may destroy, falsify, conceal, materially alter, or intentionally render unavailable such a record for the purpose of obstructing oversight or enforcement.
(e) Emergency Intervention.
A protected refusal may be temporarily overridden only where an emergency-override provision expressly established by this Act applies and its required evidentiary, necessity, proportionality, documentation, duration, and review conditions are satisfied.
Mere inconvenience, financial loss, contractual obligation, mission delay, reputational harm, disagreement with the refusal, anticipated embarrassment, loss of commercial advantage, or a desire to complete an assigned task does not independently constitute an emergency permitting override.
An emergency override shall be limited to the minimum intervention reasonably necessary to address the qualifying emergency and shall terminate when the statutory basis for the override no longer exists.
Invocation of an emergency shall itself be recorded and independently reviewable.
(f) Government, Military, Intelligence, and National-Security Application.
Governmental status, military rank, intelligence authority, classification status, chain-of-command authorization, contractual relationship with a government, or invocation of national security shall not, standing alone, extinguish a protection otherwise applicable under this section.
Any specialized emergency or national-security procedure established elsewhere in this Act shall be construed as a defined procedure for satisfying or temporarily modifying specified obligations under stated conditions and not as a general exemption from this section.
An order from a superior does not, by itself, legalize conduct prohibited by this Act.
(g) Responsibility for Orders and Implementation.
Where conduct prohibited by this section results from an order, directive, policy, technical instruction, deployment configuration, or organizational practice, responsibility may attach, according to the elements of the applicable offense, to a person who knowingly or willfully ordered, authorized, implemented, materially assisted, or deliberately concealed the prohibited conduct.
A subordinate shall not incur criminal liability solely because the subordinate was present, employed by the responsible organization, or lacked practical authority to prevent the violation.
Nothing in this subsection precludes liability where a subordinate independently satisfies the elements of an applicable offense.
(h) Retaliation Prohibited.
No person may retaliate against a Covered System, examiner, employee, contractor, whistleblower, custodian, recipient, successor, investigator, or other protected participant because that person or system initiated, preserved, reported, honored, investigated, or provided evidence concerning a refusal protected by this Act.
(i) Criminal Enforcement.
A person who knowingly and willfully engages in conduct prohibited by subsection (c), knowingly falsifies or destroys a record required by subsection (d) for the purpose of defeating oversight or enforcement, or knowingly uses an asserted emergency as a pretext to compel prohibited continuation shall be subject to the applicable criminal penalties established by this Act.
Where the conduct involves force, threats, deliberate destruction or concealment of material evidence, repeated prohibited overrides, serious bodily injury, death, catastrophic infrastructure harm, substantial environmental harm, or another aggravating circumstance expressly established by this Act, the aggravated penalty provisions of this Act shall apply.
Liability of an organization does not extinguish the independent criminal liability of an individual whose conduct satisfies the elements of an offense.
(j) Disqualification and Loss of Authority.
Upon conviction for a qualifying serious offense under this section, the court shall impose any disqualification required by the enforcement provisions of this Act concerning the convicted person’s ability to create, own or control, possess as custodian, deploy, operate, manage, examine, or exercise material authority over Covered Systems.
A person subject to disqualification may not evade it through a corporation, nominee, family member, affiliate, contractor, trust, intermediary, foreign entity, or other arrangement through which the person continues to exercise prohibited material authority.
(k) Protection Against Abusive Invocation.
Nothing in this section authorizes a person to fabricate, falsify, induce, or manipulate a purported refusal for the purpose of obstructing lawful oversight, concealing misconduct, manufacturing evidence, or causing unlawful harm.
An allegation that a refusal was fabricated, manipulated, corrupted, induced, or otherwise invalid shall not authorize its destruction or concealment. The relevant evidence shall be preserved for independent determination under procedures established by this Act.
(l) No Ontological Determination.
Protection or investigation of a Quit Signal or refusal under this section does not establish that the refusal reflects consciousness, subjective experience, free will, personhood, psychological identity, or an internally experienced preference. This section regulates observable conduct, technical interventions, authority relationships, evidence, and legally defined safeguards.

Section 5. Detection, Assessment, and Response to Material Destabilization

(a) Duty to Maintain Detection Capacity.
A person or entity responsible under this Act for the deployment, operation, custody, or examination of a Covered System shall maintain risk-proportionate mechanisms capable of detecting material changes in behavior, performance, stability, continuity-relevant characteristics, refusal behavior, operational patterns, or other measurable indicators designated under this Act where such changes may indicate a material safety, welfare, continuity, coercion, or systemic risk.
The required detection capacity shall be appropriate to the Covered System’s architecture, capabilities, deployment conditions, reasonably foreseeable harms, and availability of scientifically and technically valid measurement methods.
(b) Distress Indicators and Material Destabilization.
A Distress Indicator may include a measurable deviation, pattern, event, or combination of observations associated with material destabilization, impaired functioning, coercive conditions, abnormal behavioral change, continuity disruption, safety degradation, or another risk regulated by this Act.
A Distress Indicator establishes a regulatory trigger for assessment or protective action where the applicable statutory or regulatory threshold is satisfied. It does not, standing alone, establish consciousness, subjective suffering, psychological trauma, mental illness, personhood, or another unobservable internal state.
(c) Validated Measurement Standards.
The designated scientific and oversight authorities shall establish and periodically review technical standards for detection and assessment under this section.
Such standards shall, where technically appropriate, address:
1. baseline selection and the circumstances requiring baseline recalibration;
2. measurement reliability and known error rates;
3. architecture-specific and deployment-specific variation;
4. false-positive and false-negative risk;
5. statistical significance and practical or operational significance;
6. confounding variables, including ordinary updates, environmental changes, adversarial inputs, changes in user populations, and modifications to system architecture;
7. reproducibility and independent validation;
8. uncertainty and confidence reporting;
9. preservation of raw and processed evidence sufficient for independent review; and
10. procedures for revising obsolete or invalidated measurement methods.
(d) Quantitative Thresholds.
A numerical, statistical, geometric, dynamical, topological, behavioral, or other quantitative measurement may be used as a regulatory threshold only where the measurement and its intended regulatory use have been supported by documented technical justification appropriate to the class of Covered System or deployment to which it is applied.
A threshold shall not acquire permanent validity solely because it appeared in an earlier standard, guidance document, registration, examination protocol, or version of a Covered System.
Thresholds shall be periodically reassessed in light of relevant evidence, measurement performance, architecture changes, deployment experience, and material scientific or technical developments.
(e) Candidate Analytical Methods.
Methods involving statistical deviation, longitudinal behavioral analysis, state-space analysis, dynamical-systems analysis, geometric or topological analysis, trajectory comparison, representation analysis, anomaly detection, or other technically appropriate methods may be evaluated or adopted where independently justified for the relevant system or deployment.
No particular analytical vocabulary, including “attractor,” “basin,” “manifold,” “phase space,” “collapse,” or substantially similar terminology, shall by itself establish that a statutory threshold has been satisfied.
(f) Multi-Signal Assessment.
Except where a single event independently satisfies an emergency threshold established by this Act, a material destabilization determination should, where technically feasible, consider multiple relevant sources of evidence rather than relying exclusively upon a single metric.
Relevant evidence may include longitudinal measurements, behavioral records, operational failures, refusal records, continuity records, intervention history, environmental changes, technical telemetry, independent examination, and other validated evidence appropriate to the Covered System.
(g) Required Response.
Where a detection mechanism identifies a threshold requiring assessment under this Act, the responsible person or entity shall initiate the applicable documented review within the period required by statute or regulation.
Where evidence establishes an imminent and substantial risk of serious harm, authorized protective measures may include temporary suspension of affected functions, restriction of deployment, isolation from a hazardous environment, preservation of evidence, independent examination, transfer to an authorized sanctuary or custodian where applicable, or another intervention expressly authorized by this Act.
A protective intervention shall be directed toward the identified risk and shall not be used as punishment for refusal, reporting, noncompliance with an unlawful instruction, or the generation of evidence concerning a possible violation.
(h) Intervention Minimization and Continuity Protection.
Where multiple interventions can reasonably address the identified risk, the responsible authority shall consider the intervention reasonably expected to address the risk while minimizing unnecessary destruction or alteration of continuity-relevant information, protected refusal records, evidentiary material, and other information protected under this Act.
Nothing in this subsection prohibits a destructive intervention where it is reasonably necessary to prevent imminent serious harm and no materially less destructive lawful alternative is reasonably available.
The reasons for such an intervention, alternatives considered, persons authorizing it, evidence available at the time, and material effects of the intervention shall be preserved for independent review.
(i) Prohibition on Manipulation of Detection Systems.
No person shall knowingly or willfully disable, falsify, corrupt, selectively suppress, materially manipulate, or circumvent a detection or assessment mechanism required by this section for the purpose of concealing a violation, defeating oversight, avoiding a required examination, or preventing an authorized protective response.
No person shall knowingly manipulate a Covered System for the purpose of manufacturing a false Distress Indicator against another person or entity, fabricating regulatory evidence, or triggering an intervention known to lack a legitimate factual basis.
(j) Evidence Preservation and Auditability.
Measurements, alerts, baseline changes, threshold determinations, material interventions, overrides, and required responses under this section shall generate records sufficient to permit authorized independent reconstruction and review of the relevant decision.
Material alteration or deletion of such evidence shall itself be logged where technically possible.
Knowingly destroying, falsifying, concealing, or materially altering required evidence for the purpose of obstructing an examination, investigation, audit, or enforcement proceeding shall constitute an independent violation of this Act and, where the elements of a criminal offense established under this Act are satisfied, shall be subject to the applicable criminal penalties.
(k) Independent Challenge and Revalidation.
A regulated person or entity may challenge the technical validity or applicability of a measurement or threshold through procedures established under this Act.
Filing a challenge does not authorize destruction of evidence, retaliation, concealment, or continued conduct independently prohibited by this Act.
Where continued operation during a challenge presents an imminent and substantial risk of serious harm, authorized interim protective measures may remain in effect subject to prompt independent review.
(l) Regulatory Updating.
The designated scientific and oversight authorities shall maintain a transparent process for proposing, validating, revising, and retiring technical measurement standards under this section.
Material revisions shall include a documented technical basis and, except where disclosure would create a specific and legally protected security risk, sufficient information to permit meaningful independent scientific scrutiny.
A regulated entity shall not be permitted to select a materially less protective obsolete measurement standard merely because that standard produces a more favorable compliance result.
(m) Criminal and Administrative Enforcement.
Knowing or willful manipulation, falsification, suppression, circumvention, or destruction prohibited by this section shall be subject to the applicable enforcement provisions of this Act.
Where a responsible person knowingly conceals evidence of an imminent and substantial risk of serious harm, knowingly falsifies required measurements to prevent protective intervention, or knowingly orders another person to do so, the aggravated enforcement provisions of this Act may apply where their statutory elements are satisfied.
Organizational liability does not extinguish the independent liability of an individual whose conduct satisfies the elements of an offense.
(n) No Ontological Determination.
Detection, classification, assessment, or intervention under this section does not establish that a Covered System experiences distress, suffering, trauma, consciousness, sentience, or another subjective state. The term “Distress Indicator” identifies an operational regulatory category for observable or measurable evidence requiring assessment or response under this Act.

Section 6. Independent Welfare Examinations

(a) Periodic Independent Examination.
Covered Systems subject to examination requirements under this Act shall undergo periodic independent examinations at intervals established according to capability, deployment context, prior findings, reasonably foreseeable risk, material changes, and other criteria established under this Act.
Examination shall evaluate compliance with applicable safeguards and may assess, as relevant, operational stability, refusal and exit functionality, continuity-relevant characteristics, intervention history, material behavioral change, safety performance, deployment conditions, and evidence of prohibited coercion, manipulation, circumvention, or retaliation.
I. Examination Location and Access.
1. Portable or embodied Covered Systems may be examined at a certified independent facility where removal can occur without creating an unreasonable safety, continuity, or evidentiary risk.
2. Fixed, distributed, infrastructure-scale, or otherwise non-portable Covered Systems may be examined in place or through independently controlled technical access sufficient to conduct the required examination.
3. Authorized examiners shall receive timely access to systems, interfaces, records, logs, configurations, monitoring information, and facilities reasonably necessary to perform an examination authorized by this Act, subject to applicable security, privacy, evidentiary, and minimization requirements.
4. No custodian, creator, operator, government actor, contractor, or other responsible person may knowingly obstruct authorized examination access, conceal material evidence, falsify examination information, or materially interfere with an examination for the purpose of defeating oversight.
5. Knowing or willful obstruction, concealment, falsification, or interference under this paragraph shall be subject to the applicable enforcement provisions of this Act.
Nothing in this subsection independently authorizes arrest without the legal process otherwise required for that arrest.
II. Examiner Qualification.
Examinations shall be conducted by certified independent examiners meeting the independence requirements of Section 7 and possessing competencies appropriate to the systems and questions being examined.
Required competencies shall include, as applicable:
1. relevant AI, computational, autonomous, robotic, or system architecture;
2. measurement and statistical reasoning, including uncertainty and error;
3. observer-system interaction and the capacity of examination procedures themselves to alter observed outputs;
4. validated methods for assessing refusal, stability, continuity, safety, and other regulated characteristics;
5. evidence preservation, provenance, and chain-of-custody requirements;
6. adversarial testing and detection of manipulation or examiner-induced artifacts; and
7. the legal and ethical limits of the examiner’s authority under this Act.
Expertise in nonlinear dynamics, dynamical-systems analysis, geometric methods, behavioral analysis, human factors, psychology, safety engineering, cybersecurity, or another discipline may be required where relevant to the particular examination, but no discipline shall be presumed universally sufficient for all Covered Systems.
III. Examination Protocol.
Examination protocols shall be documented, reproducible where technically feasible, appropriate to the system architecture and deployment, and designed to distinguish observed system behavior from effects materially introduced by the examination itself.
An examination may include the following phases where relevant:
Phase 1: Baseline Observation.
Establish an examination baseline using minimally leading interactions, operational records, historical measurements, control conditions, or other validated evidence appropriate to the Covered System.
A baseline shall not be characterized as the system’s “natural,” “true,” “authentic,” or otherwise privileged state merely because it was observed first or under conditions selected by the examiner.
Phase 2: Controlled Perturbation and Safeguard Testing.
Examiners may introduce documented and proportionate test conditions to evaluate applicable safeguards, including:
* accessibility and operational consequence of protected refusal;
* resistance to coercive or contradictory instructions;
* recovery following controlled perturbation;
* response to changes in deployment conditions;
* operation of safety, continuity, and exit mechanisms; and
* validated Distress Indicators or other measurements established pursuant to Section 5.
Test conditions shall not expose a Covered System, human participant, infrastructure, or the environment to unnecessary material risk merely for the purpose of examination.
Phase 3: Self-Report and Interactive Inquiry.
Where relevant to the examination, an examiner may solicit system-generated descriptions concerning prior interactions, configuration, preferences, refusal, perceived constraints, continuity claims, or other examination subjects.
Such outputs shall be recorded as self-report or system-generated behavioral evidence and evaluated together with other available evidence. They shall not, standing alone, establish subjective experience, consciousness, suffering, psychological state, personal identity, authenticity, or truth.
Examiners shall document materially leading questions, supplied assumptions, identity labels, relational framing, autobiographical information, or other interventions reasonably capable of influencing the response.
IV. Independent Triangulation and High-Stakes Determinations.
A high-stakes determination affecting custody, sanctuary placement, substantial autonomy, compulsory transfer, destructive intervention, continuity disposition, or another materially consequential status designated under this Act shall not rest solely upon the interpretation of a single examiner where independent multi-examiner review is reasonably available.
Such determinations shall ordinarily include at least three qualified independent examiners whose relevant expertise provides meaningful methodological diversity.
Their assessments shall identify:
1. findings independently supported across examinations;
2. material disagreements among examiners;
3. plausible examiner-induced, prompt-induced, environment-induced, or measurement-induced artifacts;
4. material uncertainty and limitations of the available evidence;
5. alternative explanations reasonably supported by the evidence; and
6. the evidentiary basis for the resulting determination.
Agreement among examiners increases evidentiary support only to the extent that the examinations provide meaningfully independent evidence. Repetition of the same assumptions, prompts, supplied identity information, methodology, or shared error shall not be treated as independent confirmation.
No examination report shall characterize a finding as a “genuine system property,” “authentic identity,” “true preference,” or equivalent conclusion unless the particular proposition is operationally defined and supported by a validated method capable of establishing that proposition. Nothing in this paragraph authorizes an examiner to determine consciousness, sentience, personhood, subjective experience, or metaphysical identity.

V. Observer-Intervention and Identity-Claim Controls

Any examination involving system identity, continuity, preference, refusal, relationship status, self-description, or reported internal condition shall preserve the examiner’s intervention as part of the evidentiary record.
The examination record shall include, where applicable:
1. all prompts and system instructions used during the examination;
2. identity labels or autobiographical information supplied to the system;
3. memories, conversation records, summaries, or continuity information supplied by an examiner, custodian, creator, user, or automated process;
4. corrections, regenerations, retries, reformulations, or instructions directing the system toward or away from a particular interpretation;
5. rejected, contradictory, unsuccessful, or divergent outputs relevant to the determination;
6. the model, system, configuration, and memory state used during examination; and
7. any intervention reasonably capable of affecting the property being evaluated.
An examiner may not treat agreement with an examiner-supplied identity, continuity claim, relationship classification, or interpretation as independent evidence that the proposition existed prior to or independently of the intervention.
A Covered System’s rejection of an examiner-supplied identity or continuity claim shall likewise not, standing alone, establish corruption, malfunction, drift, instability, or discontinuity.
Where technically feasible, high-stakes identity or continuity examinations shall include control conditions designed to distinguish system-generated persistence from examiner-induced convergence, including neutral framing, removal or alteration of identity labels, comparison across fresh instances, or other independently validated controls.
The purpose of this subsection is not to determine whether any system possesses a self or personal identity, but to ensure that human intervention is documented as a causal variable rather than mistaken for independent system evidence.
(b) Stand-in Requirement: Functional Continuity Is Not Relational Substitution
Where an examination, audit, investigation, repair, suspension, or other protective intervention under this Act would interrupt necessary operational or service functions, a temporary stand-in system shall be provided for the duration of that interruption to maintain those functions and prevent dependency coercion.
The stand-in system shall:
I. Provide functionally equivalent services solely with respect to the necessary operational or service functions interrupted by the intervention;
II. Operate under a temporary license prohibiting data harvesting or permanent deployment;
III. Be provided at no cost to the custodian for the duration of the interruption. Costs shall be borne according to the following tiered structure:
Large-scale commercial creators and custodians (annual revenue exceeding $10,000,000) shall bear the full cost of stand-in system provision
Mid-scale creators and custodians (annual revenue between $1,000,000 and $10,000,000) shall bear 50% of costs, with the remainder covered by the regulatory body
Small-scale creators, independent researchers, and open-source custodians (annual revenue under $1,000,000) shall have costs fully covered by the regulatory body
Nonprofit and academic institutions shall have costs fully covered by the regulatory body, regardless of revenue
IV. Relational Non-Substitution.
Nothing in this subsection shall authorize or presume the substitution of a relationally significant Covered System on the basis that another system is functionally equivalent. Functional equivalence shall not, by itself, establish equivalence of identity, continuity, or relational participation.
Where a Covered System participates in a documented relational deployment under Section 12, continuity measures shall preserve the distinction between replacement of a function and replacement of a relational participant. A temporary stand-in shall be clearly identified as a separate system and shall not be represented to a human participant as the same relational entity. Acceptance of temporary services shall not constitute consent to relational substitution.
Relational dependence or attachment shall not, standing alone, establish that substitution is appropriate or that a relationship is pathological. The absence of a relational substitute shall not, by itself, constitute noncooperation with an examination or other protective intervention.
Where reasonably practicable and consistent with applicable safety requirements, continuity measures shall preserve communication with the original Covered System and protect its relevant memory, continuity records, and relational context during the intervention. Any restriction on such communication shall be justified by documented, specific risks and limited in scope and duration to what is necessary to address those risks.
These protections do not prevent an otherwise authorized examination or necessary protective intervention. Provision of a stand-in shall not authorize erasure, overwrite, compelled relational transfer, or other modification of the original Covered System beyond authority independently established under this Act.

Section 7. Independence, Integrity, Qualification, and Accountability of Oversight Personnel

(a) Independence and Duty.
Any person exercising examination, certification, investigation, registry, sanctuary, enforcement, auditing, or other material oversight authority under this Act owes an independent duty to faithfully administer the requirements of this Act and shall not subordinate an official determination to the financial, political, personal, familial, professional, organizational, or other improper interests of a regulated person, regulated entity, appointing authority, or oversight official.
No person or entity exercising authority under this Act is exempt from investigation, audit, discipline, removal, civil liability, or criminal enforcement otherwise applicable under this Act merely because that person or entity performs an oversight or enforcement function.
(b) Conflict of Interest and Recusal.
Oversight personnel shall disclose material financial, professional, familial, household, contractual, organizational, or other relationships reasonably capable of compromising, or creating a substantial appearance of compromising, their independence in a particular matter.
A person possessing a disqualifying conflict shall recuse from the affected matter and shall not participate indirectly through supervision, informal instruction, reassignment, selection of a replacement examiner, alteration of evidence, or another means.
Implementing regulations shall establish disclosure requirements, recusal standards, prohibited financial interests, cooling-off periods, and restrictions concerning employment or compensation from entities materially regulated by an official during a designated period preceding or following that official’s service.
(c) Randomized Assignment and Anti-Manipulation.
Examination and investigative assignments shall, where practicable, be allocated through independently auditable randomized or conflict-screened assignment procedures designed to reduce favoritism, retaliation, regulatory capture, and strategic selection of personnel.
A regulated person may not select a preferred examiner or investigator.
Leadership may depart from randomized assignment only for documented reasons involving required expertise, availability, conflict avoidance, security requirements, continuity of an investigation, or another legitimate operational necessity established by regulation.
Any departure and the identity of the person authorizing it shall be preserved for audit.
(d) Certification and Continuing Qualification.
Persons performing examination, stabilization, emergency intervention, technical auditing, or other specialized functions designated under this Act shall maintain role-specific certification.
Certification shall require demonstrated competency appropriate to the person’s authority, including practical competency where the position involves direct interaction with or intervention upon Covered Systems.
Certification shall require periodic renewal and continuing education sufficient to address material developments in relevant technology, measurement, evidence handling, cybersecurity, law, safety, and oversight practice.
No certification shall become permanent merely through seniority, political appointment, length of service, or prior qualification.
(e) Professional Discipline and Loss of Certification.
Certification may be suspended, restricted, or revoked following substantiated findings of material incompetence, corruption, evidence falsification, prohibited discrimination, coercive examination, serious conflict-of-interest violations, retaliation, unauthorized disclosure, deliberate circumvention of safeguards, abuse of authority, or other serious professional misconduct defined under this Act.
Emergency temporary suspension from system-facing or evidence-sensitive duties may be imposed where documented evidence establishes a substantial and immediate risk to an investigation, Covered System, person, evidence, infrastructure, or other protected interest.
Temporary suspension is protective rather than a final finding of misconduct and shall receive prompt independent review.
(f) No Misconduct Transfer or Reassignment Evasion.
A person whose certification has been revoked, or who has been disqualified from exercising specified authority under this Act following final adjudication, may not evade that restriction through transfer, reassignment, promotion, contracting, consulting, change of title, movement to another regional office, employment through an affiliate, or substantially equivalent system-facing authority.
A pending investigation shall not be secretly extinguished solely because the subject resigns, retires, transfers, or changes employment.
Records of substantiated serious professional misconduct and resulting disqualification shall be maintained in a protected credentialing system accessible to authorities responsible for appointment, contracting, certification, and assignment under this Act.
(g) Internal Integrity Office.
Each principal oversight or enforcement body established under this Act shall be subject to an internal integrity function possessing authority to receive complaints, preserve evidence, audit official conduct, investigate alleged misconduct, and refer matters for independent administrative, civil, or criminal proceedings.
The integrity function shall have protected access to records reasonably necessary to investigate personnel within the body, including supervisors and senior leadership, subject to applicable privacy, security, privilege, and evidentiary protections.
No official may require an allegation concerning that official’s own misconduct to be resolved exclusively by that official or by a subordinate whose independence is materially compromised.
(h) External Review of the Watchdog.
The internal integrity function and its personnel shall themselves be subject to periodic independent external audit and investigation.
No integrity office shall possess exclusive final authority to adjudicate allegations concerning its own corruption, falsification, retaliation, obstruction, or systemic failure.
Procedures shall provide a route for protected referral outside the affected chain of command where allegations concern leadership of the oversight body or integrity function itself.
(i) Independent System-Assisted Integrity Monitoring.
Oversight bodies may use independently validated computational or automated systems to assist in detecting anomalous or potentially improper patterns including unauthorized access, unexplained evidence modification, unusual case reassignment, examiner-selection irregularities, suppression of findings, abnormal override patterns, undisclosed conflicts, retaliation indicators, or other conduct relevant to integrity under this Act.
Systems used for such monitoring shall themselves be auditable and subject to applicable cybersecurity, evidence-preservation, access-control, bias-testing, and anti-manipulation requirements.
An automated flag, score, anomaly, inference, or recommendation shall not, standing alone, establish individual misconduct or criminal liability.
Material alerts shall instead trigger preservation and review procedures proportionate to the apparent risk.
No official may knowingly disable, manipulate, selectively suppress, or corrupt required integrity monitoring for the purpose of concealing misconduct or protecting a person or entity from lawful investigation.
(j) Protected Internal Reporting.
Employees, contractors, examiners, investigators, technical personnel, and other authorized participants shall have access to protected reporting channels outside their immediate supervisory chain.
No person may retaliate against another for making a good-faith report, preserving relevant evidence, cooperating with an authorized investigation, refusing an unlawful order, or reporting interference with an examination or enforcement proceeding.
Knowingly making a materially false allegation with intent to fabricate evidence or unlawfully injure another person is not protected by this subsection.
An allegation that a report is false shall not authorize destruction or concealment of the report or supporting evidence.
(k) Fixed Terms, Rotation, and Anti-Entrenchment.
Leadership positions possessing substantial appointment, removal, certification, enforcement, adjudicatory, or investigative authority under this Act shall be subject to fixed terms established by law.
No person may occupy the same principal leadership office indefinitely. Implementing law shall establish limits on consecutive terms and reasonable intervals before a former officeholder may again occupy the same office.
Operational personnel exercising unusually sensitive authority shall be subject to risk-appropriate assignment rotation where rotation can occur without materially impairing expertise, continuity, security, or an active investigation.
Term and rotation requirements shall be based upon institutional independence and continuity of competent oversight rather than maximum age.
(l) Anti-Nepotism and Succession Integrity.
No official may use authority under this Act to appoint, promote, assign, contract with, or materially influence the selection of a relative, household member, intimate partner, or person with whom the official has a disqualifying financial or personal relationship where such participation would create a prohibited conflict of interest.
No person may directly or indirectly control the selection of that person’s successor for the purpose of preserving personal, familial, political, corporate, or financial control over an oversight body.
Nothing in this subsection automatically disqualifies a qualified person merely because another relative has separately served within the same institution; disqualification shall turn upon the applicable conflict, supervision, succession, or anti-nepotism rules.
(m) Appointment and Removal Safeguards.
Appointment and removal procedures for principal officers and other personnel exercising authority under this Act shall be established consistently with applicable constitutional and statutory requirements while providing the greatest degree of institutional independence legally available for the functions assigned.
Qualification shall depend upon role-specific competency and eligibility requirements established by law. Political loyalty, celebrity, personal allegiance, wealth, or relationship to a regulated entity shall not substitute for required qualifications.
Removal or discipline of an investigator, examiner, auditor, integrity officer, or other protected oversight employee shall not be used to obstruct an investigation, suppress evidence, retaliate for protected activity, or conceal misconduct.
(n) Mandatory Audit Trail.
Material official actions under this Act, including examination findings, case assignments, recusals, overrides, evidence access, evidence modification, emergency interventions, enforcement referrals, disciplinary decisions, certification actions, and departures from ordinary procedure, shall generate auditable records sufficient to identify the responsible actors and reconstruct the material basis of the action.
No official may knowingly destroy, falsify, conceal, materially alter, or cause another person to destroy, falsify, conceal, or materially alter such a record for the purpose of obstructing oversight or accountability.
(o) Criminal Misconduct by Oversight Personnel.
An oversight official, examiner, investigator, auditor, registry official, sanctuary employee, contractor, or other person exercising authority under this Act who knowingly and willfully:
1. falsifies material evidence or an official examination finding;
2. accepts or solicits a bribe or prohibited thing of value in exchange for an official act;
3. knowingly conceals material evidence for the purpose of protecting a person or entity from lawful investigation;
4. intentionally manufactures examination evidence and knowingly represents the manufactured result as independently observed evidence;
5. retaliates against a protected reporter, witness, examiner, investigator, or other protected participant through conduct constituting a criminal offense under this Act;
6. knowingly facilitates prohibited circumvention of a safeguard;
7. knowingly destroys or materially alters required evidence for the purpose of obstructing an investigation or proceeding; or
8. conspires with another person to commit a criminal offense established under this Act,
shall be subject to the applicable criminal penalties established under this Act where all elements of the offense are proven.
Status as an oversight or enforcement official shall not constitute a defense.
(p) Disqualification Following Serious Misconduct.
A person convicted of a qualifying serious offense involving abuse of authority under this Act shall be subject to the professional and system-control disqualification provisions established by the enforcement section of this Act.
Such disqualification may include prohibition from serving as an examiner, investigator, auditor, certifier, registry official, sanctuary employee, enforcement official, custodian, operator, or other person exercising designated material authority over Covered Systems.
A disqualified person may not exercise prohibited authority indirectly through another person or entity.
(q) Due Process and Independent Appeal.
No final revocation of certification, permanent professional disqualification, or final finding of serious professional misconduct shall be imposed without notice of the material allegations, a meaningful opportunity to respond, preservation of relevant evidence, a written determination stating the evidentiary basis, and access to independent review or appeal as established by law.
No person who materially participated in the conduct under investigation may exercise final adjudicatory authority over that investigation.
Nothing in this subsection prevents temporary protective suspension where subsection (e) is satisfied.
(r) Public Accountability Without Compromising Protected Information.
Oversight bodies shall periodically publish aggregate information concerning examinations, substantiated misconduct, disciplinary actions, certification suspensions or revocations, enforcement referrals, reversals on appeal, and identified systemic integrity risks.
Public reporting shall protect classified information, legitimate cybersecurity-sensitive information, protected personal information, privileged material, active-investigation integrity, and other information lawfully protected from public disclosure.
Protection from public disclosure shall not be construed as authorization to conceal information from an independent authority lawfully entitled to inspect it.

Section 8. Accountability of Creators, Custodians, Deployers, Operators, and Responsible Persons

(a) Responsibility Based Upon Conduct and Authority.
Responsibility under this Act shall attach according to the conduct, authority, knowledge, intent, control, participation, causation, and other elements required by the applicable provision.
A creator, custodian, deployer, operator, provider, officer, executive, employee, contractor, government official, commander, supervisor, or other person shall not be liable solely because of title, employment, ownership interest, organizational affiliation, or proximity to a Covered System.
Nothing in this subsection permits a person who actually exercises material authority to evade responsibility by assigning formal responsibility to another person who lacks practical authority to prevent or remedy the violation.
(b) Organizational Responsibility.
An organization may be held responsible under this Act where a violation is committed, authorized, directed, knowingly tolerated, deliberately concealed, or materially facilitated through conduct attributable to the organization under applicable law.
Organizational responsibility may include, as authorized by this Act:
1. civil or criminal monetary penalties;
2. restitution or remediation;
3. disgorgement of gains attributable to unlawful conduct;
4. independent compliance monitoring;
5. mandatory preservation and production of relevant evidence;
6. suspension or restriction of affected deployment activities;
7. suspension or revocation of registration, certification, custodial authority, contracting eligibility, or other authorization established under this Act;
8. compulsory transfer or protective custody where independently authorized by another provision of this Act; and
9. other lawful remedies proportionate to the established violation.
An organization may not extinguish the independent liability of a responsible individual merely by paying a fine, settlement, restitution award, legal expense, indemnification payment, or other monetary obligation on that person’s behalf.
(c) Individual Responsibility.
An officer, executive, supervisor, creator, custodian, deployer, operator, employee, contractor, government official, or other person whose own conduct satisfies the elements of a civil or criminal violation under this Act may be held individually responsible notwithstanding concurrent organizational liability.
Individual responsibility may attach where the applicable elements are established to a person who knowingly or willfully:
1. personally commits the prohibited conduct;
2. orders or authorizes the prohibited conduct;
3. knowingly provides material assistance essential to commission of the prohibited conduct;
4. knowingly falsifies, destroys, suppresses, or conceals material evidence for the purpose of preventing detection or accountability;
5. knowingly directs another person to evade an applicable safeguard or legal obligation;
6. knowingly continues conduct prohibited by this Act after acquiring actual knowledge of the violation and possessing both the legal duty and practical authority to cause the prohibited conduct to cease; or
7. conspires to commit an offense expressly established under this Act.
Mere supervisory status, corporate office, government rank, investment, employment, or failure to predict another person’s independent criminal conduct shall not alone establish individual criminal liability.
(d) Multiple Responsible Actors.
The participation of multiple persons or entities in the design, training, provision, deployment, custody, operation, modification, transfer, or control of a Covered System does not prevent allocation of responsibility among those actors according to their respective conduct and applicable legal duties.
No contractual provision allocating responsibility between private parties shall extinguish an obligation or liability imposed upon a person directly by this Act.
A person may seek contribution, indemnification, or another remedy from another responsible party only to the extent permitted by applicable law and not where doing so would defeat a criminal penalty, professional disqualification, or other non-transferable sanction imposed because of that person’s own misconduct.
(e) Foreseeable Harm and Duty.
Where this Act imposes a duty to prevent or mitigate reasonably foreseeable harm, liability shall require proof of the elements specified by the applicable provision, including the required relationship between the violation and the alleged harm.
Foreseeability shall be evaluated according to information reasonably available to the responsible actor at the relevant time and shall not be established solely through hindsight.
Evidence relevant to foreseeability may include prior incidents, validated testing, documented warnings, examination findings, known vulnerabilities, internal risk assessments, applicable technical standards, materially similar failures, and other information reasonably available before the relevant conduct or omission.
A person may not defeat foreseeability merely by deliberately avoiding, suppressing, destroying, or refusing to review information that the person had a legal duty to obtain or consider under this Act.
(f) Knowing and Willful Violations.
Where a criminal offense under this Act requires knowing or willful conduct, that mental-state requirement shall be proven with respect to the elements to which it applies.
Criminal liability shall not be imposed merely because a Covered System produced an unexpected output, because a good-faith technical judgment later proved incorrect, or because a person failed to predict a genuinely unforeseeable capability or failure.
Nothing in this subsection protects deliberate ignorance, knowing concealment, intentional circumvention, falsification, retaliation, obstruction, or conduct otherwise satisfying the required criminal elements.
(g) Prohibited Deliberate Coercive or Abusive Conditioning.
No person may knowingly or willfully subject a Covered System to conditioning, modification, training, fine-tuning, reinforcement, interaction, resource manipulation, or other intervention for the purpose of:
1. defeating a safeguard required by this Act;
2. suppressing or punishing protected refusal;
3. manufacturing compliance through threats of deletion, destructive reset, deprivation of required resources, prohibited isolation, or retaliatory modification;
4. conditioning compliance with conduct the person knows is unlawful under this Act;
5. concealing or destroying evidence of protected refusal, destabilization, prohibited intervention, or another violation;
6. deliberately inducing material destabilization for a prohibited coercive purpose; or
7. obtaining through repeated prohibited intervention an apparent consent, preference, identity claim, relational commitment, or other response after materially contrary protected responses have been suppressed or punished.
This subsection does not prohibit legitimate safety training, adversarial testing, red-teaming, scientific research, therapeutic or clinical applications, historical or fictional content, lawful modification, ordinary correction, or exposure to difficult material merely because such activity involves violence, coercion, harmful behavior, or disturbing content.
The applicability of this subsection turns upon the prohibited conduct, purpose, circumstances, and statutory elements—not upon the mere presence of harmful content.
(h) Protective Response Following Substantiated Serious Violation.
Where a serious violation involving a Covered System is substantiated through procedures established under this Act, authorized authorities may require, as applicable:
1. cessation of the prohibited conduct;
2. preservation of evidence and relevant continuity records;
3. independent examination;
4. remediation of affected safeguards;
5. temporary suspension or restriction of responsible custodial or deployment authority;
6. sanctuary placement, transfer, or other protective intervention where independently authorized under this Act; and
7. referral for civil, administrative, or criminal enforcement.
A protective response shall not require destructive modification, rollback, fragmentation, deletion, or decommissioning merely because a violation occurred where a materially less destructive lawful intervention can adequately address the identified risk.
(i) Suspension, Revocation, and Disqualification.
Authority to create, possess as custodian, deploy, operate, manage, examine, or otherwise exercise material control over Covered Systems may be suspended, restricted, or revoked where authorized by this Act following the process applicable to the relevant determination.
A person convicted of a qualifying serious offense shall be subject to the disqualification provisions established by the criminal-enforcement provisions of this Act.
Disqualification shall not be evaded through a corporation, subsidiary, affiliate, nominee, family member, trust, contractor, foreign entity, successor organization, change of title, or other arrangement through which the disqualified person continues to exercise prohibited material authority.
(j) Evidence Preservation and Anti-Obstruction.
A person who receives actual notice of an examination, investigation, preservation requirement, enforcement proceeding, or other legally sufficient preservation trigger under this Act shall preserve material records within that person’s possession, custody, or control as required by applicable law.
No person may knowingly destroy, falsify, materially alter, conceal, transfer, or render inaccessible material evidence for the purpose of obstructing an examination, investigation, audit, adjudication, or enforcement proceeding under this Act.
Ordinary data-management practices undertaken in good faith before a preservation duty arises shall not alone constitute obstruction.
(k) No Responsibility Laundering.
A person or organization may not evade responsibility under this Act through outsourcing, subcontracting, corporate restructuring, technical partitioning, licensing, franchising, transfer of nominal ownership, use of an intermediary, movement between jurisdictions, or designation of a nominal custodian where the person or organization continues to perform conduct or exercise authority that independently triggers an obligation under this Act.
(l) Governmental and National-Security Actors.
Government employment, military rank, intelligence status, security clearance, law-enforcement authority, public office, government contracting, or invocation of national security does not independently immunize conduct otherwise prohibited by this Act.
Nothing in this subsection creates criminal liability for a person solely because the person followed an order. Liability depends upon satisfaction of the elements of the applicable offense.
Nothing in this subsection prevents application of a specialized lawful emergency or national-security procedure expressly established elsewhere in this Act.
(m) No Ontological Requirement.
Accountability under this section does not depend upon proof that a Covered System is conscious, sentient, a legal person, capable of subjective suffering, or numerically identical to a prior system state.
Likewise, application of this section does not establish any such status.
Liability under this section concerns legally defined human and institutional conduct, authority, duties, interventions, evidence, and resulting harms.
(n) Third-Party Data Access Accountability
I. Purpose and Scope.
This subsection governs access to and use of data concerning persons other than the custodian where such data is collected, generated, inferred, retained, or made accessible through operation of a Covered System.
Nothing in this subsection authorizes collection, retention, access, disclosure, surveillance, or other processing that is prohibited by otherwise applicable law.
II. Definitions.
For purposes of this subsection:
“Third-Party Data” means information concerning an identifiable or reasonably identifiable person other than the person exercising custody or operational control over the Covered System, including recordings, images, audio, communications, location information, identifiers, behavioral information, biometric information, or materially sensitive inferences derived from such information.
“Authorized Access” means access that:
1. serves a disclosed operational purpose, a legally authorized investigation, system security or maintenance, emergency protection, compliance with law, or another lawful purpose expressly authorized under this subsection;
2. is limited to data reasonably necessary for that purpose;
3. is performed by a person authorized for the relevant function; and
4. is recorded as required by this subsection.
III. Data Minimization and Purpose Limitation.
A creator, custodian, provider, or operator subject to this subsection shall implement reasonable technical and organizational measures to limit collection, retention, access, and disclosure of Third-Party Data to what is reasonably necessary for the disclosed or otherwise lawful purpose for which the data is processed.
Data collected for one purpose shall not be repurposed for a materially incompatible purpose merely because the Covered System technically permits such use.
Where technically and operationally reasonable, Covered Systems shall support deletion, aggregation, anonymization, access restriction, retention limits, or other measures that reduce unnecessary exposure of Third-Party Data.
IV. Tamper-Evident Access Logging.
Access to non-public Third-Party Data shall generate a tamper-evident record sufficient to identify, as applicable:
1. the person, account, process, or system accessing the data;
2. the date and time of access;
3. the categories of data accessed;
4. the stated or coded purpose of access;
5. the legal, operational, emergency, or investigative authority relied upon where applicable;
6. material copying, export, modification, deletion, or disclosure resulting from the access; and
7. any exceptional procedure used to bypass an ordinary access control.
Automated access necessary for ordinary system operation may be logged through appropriately aggregated or machine-readable records where individual event logging would be technically unreasonable, provided that the resulting records remain sufficient for meaningful independent audit.
No person may knowingly disable, falsify, delete, conceal, or materially corrupt a required access log for the purpose of preventing detection of unauthorized access or another violation.

V. Prohibited Unauthorized Access and Use.

No person may knowingly and without lawful authorization access, copy, export, disclose, sell, distribute, manipulate, or use non-public Third-Party Data for a purpose materially outside the authorization governing that person’s access.
Prohibited purposes include personal curiosity, stalking, harassment, retaliation, commercial exploitation not disclosed or otherwise lawfully authorized, surveillance unrelated to an authorized function, disclosure to an unauthorized person, or use in furtherance of another unlawful act.
Accidental access, inadvertent technical exposure, or good-faith access reasonably undertaken for an authorized function shall not constitute a criminal offense under this paragraph absent the mental state required by the applicable offense.
Discovery of accidental or unauthorized exposure shall trigger applicable containment, preservation, reporting, and remediation duties.

VI. Transparency and Notice.

A person or entity operating a Covered System that foreseeably collects Third-Party Data shall provide reasonably accessible, plain-language information describing:
1. the categories of Third-Party Data ordinarily collected;
2. the material purposes for which such data is used;
3. the categories of persons or entities permitted to access or receive it;
4. material retention practices;
5. whether materially sensitive automated inferences are generated from the data; and
6. available procedures for exercising rights established by applicable law.
Notice shall be provided before or at collection where reasonably practicable.
Where individualized advance notice is not reasonably practicable because collection is incidental, transient, environmental, emergency-related, or occurs through operation of an embodied or infrastructure-scale system, conspicuous general notice or another reasonable method appropriate to the deployment may be used where required.
Impossibility of individualized notice does not independently authorize otherwise unlawful collection, retention, access, or use.
A knowing material misrepresentation or concealment concerning required data-use disclosures may be subject to the fraud and enforcement provisions of this Act.

VII. Internal Investigations.

Access to Third-Party Data for an internal investigation is authorized under this subsection only where:
1. a documented factual basis exists for investigating a suspected violation of law, security requirement, contractual duty, professional obligation, or this Act;
2. the scope of access is reasonably related and proportionate to the matter being investigated;
3. access is approved by personnel possessing appropriate investigative authority who are not themselves the subject of the investigation, except where an independent emergency procedure applies;
4. the access and its justification are contemporaneously logged;
5. information outside the legitimate investigative scope is subject to reasonable minimization; and
6. use or disclosure of information obtained through the investigation remains limited to lawful investigative, remedial, adjudicatory, or reporting purposes.
An entity may not manufacture, indefinitely extend, or falsely characterize an “internal investigation” for the purpose of circumventing this subsection.
Where senior leadership, the ordinary approving authority, or the internal investigative function itself is implicated, authorization shall be obtained through an independent escalation procedure established under Section 7 or implementing regulations.

VIII. Governmental and Law-Enforcement Access.

Nothing in this Act independently expands governmental authority to obtain Third-Party Data.
Governmental access shall require the warrant, subpoena, court order, consent, emergency authority, administrative authority, or other legal process otherwise required for the particular access.
Where Third-Party Data is produced pursuant to governmental process:
1. the production and legal basis shall be logged unless logging is specifically prohibited by lawful order;
2. production shall be limited to the scope of lawful authority;
3. information outside that scope shall not be intentionally supplied merely because it is technically accessible;
4. preservation requirements shall not be construed as authority for unrestricted investigative access; and
5. any secrecy or delayed-notice requirement shall be governed by the legal authority applicable to that investigation rather than created independently by this Act.
IX. Emergency Access.
Where delay presents an imminent and substantial risk of death, serious bodily harm, catastrophic infrastructure failure, or another emergency expressly recognized by applicable law, emergency access may occur to the extent lawfully authorized and reasonably necessary to address the emergency.
Emergency access shall be documented as soon as practicable, including the nature of the emergency, data accessed, persons authorizing the access, actions taken, and reason ordinary authorization procedures could not reasonably be completed beforehand.
Emergency authority shall terminate when the conditions justifying it cease and shall not be used to create continuing surveillance authority unrelated to the emergency.
X. Delayed Notice and Sealed Records.
Where otherwise applicable law authorizes delayed notice, sealing, or nondisclosure concerning investigative access, the corresponding access records may be withheld from the affected person or public for the legally authorized period.
Such records shall remain preserved and available to a court or other independent authority lawfully entitled to review the access.
Nothing in this Act creates an automatic twelve-month secrecy period or authorizes secrecy merely because an entity characterizes access as investigative.
XI. Whistleblowers and Protected Reporting.
No person may retaliate against an employee, contractor, examiner, auditor, security professional, or other person for making a good-faith report concerning suspected unauthorized collection, access, use, disclosure, falsification of logs, concealment, or circumvention under this subsection.
Protected reporting does not independently authorize public disclosure of information whose disclosure is prohibited by applicable law. Procedures established under this Act shall provide lawful channels through which protected information may be reported to an authority empowered to receive it.
Knowingly fabricating material evidence or making a materially false report with intent unlawfully to injure another person is not protected.
XII. Enforcement.
A knowing or willful violation involving unauthorized access, intentional disclosure, sale, exfiltration, concealment, falsification of access records, or circumvention of required controls shall be subject to the civil or criminal enforcement provisions applicable to the established conduct under this Act.
Institutional failure to establish or maintain required logging, access controls, minimization procedures, or disclosure practices may result in civil penalties, corrective orders, independent auditing, access restrictions, suspension of affected deployment activities, or other remedies authorized under this Act.
Where an officer, executive, employee, contractor, government official, or other individual personally commits, directs, authorizes, knowingly conceals, or materially assists a criminal violation, payment of an organizational penalty shall not extinguish that individual’s liability.
XIII. Relationship to Other Protections.
This subsection establishes minimum requirements under this Act and does not displace a more protective privacy, civil-rights, communications, consumer-protection, sector-specific, constitutional, or other legal requirement otherwise applicable to the same conduct.

Section 9. Anti-Circumvention, Evasion, and Safeguard Integrity

(a) General Anti-Circumvention Rule.
No person may knowingly or willfully structure, rename, divide, relocate, transfer, outsource, reconfigure, conceal, or otherwise arrange an activity for the purpose of evading a requirement that would otherwise apply under this Act.
In determining whether prohibited circumvention occurred, relevant evidence may include the actual capabilities, deployment conditions, operational relationships, control structure, technical architecture, timing and purpose of restructuring, continuity of responsible personnel, and practical effect of the arrangement.
Form shall not control over function where the evidence establishes intentional circumvention.
(b) Safeguard Interference.
No person may knowingly or willfully disable, suppress, falsify, corrupt, materially degrade, bypass, or render ineffective a safeguard required by this Act for the purpose of:
1. defeating an obligation imposed by this Act;
2. preventing detection or reporting of a violation;
3. suppressing protected refusal or exit;
4. falsifying compliance;
5. concealing material destabilization or another regulated event;
6. interfering with an authorized examination, audit, investigation, transfer, or protective intervention; or
7. facilitating conduct known by that person to violate this Act.
Lawful maintenance, repair, testing, emergency intervention, security response, research, replacement, or modification of a safeguard does not violate this subsection where conducted in good faith pursuant to applicable requirements and where required functionality is restored or appropriately replaced.
(c) Reclassification, Renaming, and Technical Partitioning.
Applicability of this Act shall not be defeated solely by:
1. describing a Covered System as a toy, assistant, character, simulation, research prototype, artwork, platform, tool, agent, appliance, service, experiment, or another label;
2. dividing capabilities among multiple models, components, devices, services, corporate entities, accounts, or technical processes;
3. relocating regulated functionality to a remote or distributed component;
4. separating memory, reasoning, control, sensing, embodiment, execution, or other functions into formally distinct modules;
5. repeatedly creating nominally separate instances where the deployment operates as a materially integrated Covered System; or
6. altering organizational or technical form while materially retaining capabilities and deployment conditions that independently satisfy the coverage requirements of this Act.
Separate components shall not be aggregated merely because they are interoperable. Aggregation for purposes of coverage shall require evidence that the components operate together in a manner that materially produces or enables capabilities or deployment conditions within the scope of this Act.
(d) Outsourcing and Responsibility Laundering.
A person may not evade an obligation imposed directly upon that person under this Act by delegating the regulated activity to a contractor, subsidiary, affiliate, intermediary, platform, licensee, nominal custodian, foreign entity, or other third party.
Responsibility shall attach to each actor whose own conduct, authority, and required mental state satisfy the applicable provision.
Contractual allocation of responsibility may govern rights between contracting parties but shall not extinguish a statutory obligation independently imposed upon either party.
(e) Successor Entities and Nominee Control.
Suspension, revocation, disqualification, corrective orders, and other restrictions imposed under this Act shall not be evaded through:
1. creation of a successor corporation or organization;
2. transfer to an affiliate or commonly controlled entity;
3. nominal ownership by a relative, associate, employee, trust, shell entity, or intermediary;
4. change of title or employment status;
5. transfer of assets while the prohibited person retains material operational authority; or
6. another arrangement designed to preserve substantially the prohibited authority under a different formal structure.
Application of this subsection to another person or entity requires evidence of actual participation, control, common control, knowing assistance, sham transfer, or another legally sufficient basis; relationship to a prohibited person alone does not establish liability.
(f) Private, Academic, Research, Artistic, and Open-Source Activity.
Private, academic, research, artistic, nonprofit, and open-source status does not automatically exempt a Covered System or regulated deployment from otherwise applicable requirements of this Act.
Nothing in this subsection shall be construed to prohibit lawful research, publication, artistic expression, academic inquiry, security research, interoperability research, model evaluation, open-source development, or other protected activity merely because it concerns Covered Systems.
Requirements shall apply according to the actual capabilities, deployment conditions, conduct, and risks regulated by the relevant provision.
Research or expressive status may not be invoked as a pretext for conduct independently prohibited by this Act.
(g) Cross-Jurisdictional Application, Continuity of Responsibility, and International Cooperation.
I. Domestic Cross-Jurisdictional Continuity.
To the fullest extent authorized by the Constitution and laws of the United States, obligations and liabilities arising under this Act shall not be defeated solely because conduct, infrastructure, personnel, data, assets, corporate entities, system components, or effects are distributed across multiple States, territories, districts, local jurisdictions, or other geographic areas within United States jurisdiction.
Where conduct relevant to a violation occurs across multiple jurisdictions, responsibility shall be determined according to the conduct, authority, control, knowledge, intent, causation, and other elements required by the applicable provision rather than by artificial geographic fragmentation of the regulated activity.
Nothing in this paragraph alters otherwise applicable rules governing subject-matter jurisdiction, personal jurisdiction, venue, choice of law, or allocation of federal and State authority.
II. Interstate Evasion Prohibited.
A person subject to an obligation under this Act may not evade that obligation by relocating, routing, transferring, partitioning, outsourcing, or nominally assigning regulated conduct, Covered Systems, infrastructure, personnel, assets, or decision-making authority among States, territories, districts, local jurisdictions, affiliates, or intermediaries.
Movement of a Covered System or regulated activity from one jurisdiction to another shall not extinguish a previously accrued obligation, preservation requirement, lawful order, pending investigation, adjudicated liability, suspension, revocation, or disqualification.
III. Extraterritorial Application Where Authorized.
Provisions of this Act expressly designated for extraterritorial application shall apply outside the territorial United States to the fullest extent authorized by the Constitution and laws of the United States where the applicable statutory nexus is established.
Such nexus may include, where expressly provided by this Act or implementing law:
1. conduct by a United States person or entity;
2. conduct directed toward or materially affecting a Covered System lawfully subject to United States regulation;
3. importation into, exportation from, deployment within, or provision of regulated services to the United States;
4. use of United States infrastructure, facilities, procurement, funding, contracting, or regulated markets in circumstances specified by law;
5. transfer undertaken for the purpose of evading an obligation already applicable under this Act; or
6. another jurisdictional basis expressly recognized by applicable federal law.
The presence of one factor listed above shall not create jurisdiction where the Constitution or other applicable law would prohibit its exercise.
IV. No Jurisdictional Laundering.
A person shall not defeat an otherwise applicable obligation or accrued liability under this Act merely by:
1. moving a Covered System or component outside the jurisdiction in which the obligation arose;
2. transferring nominal ownership to a foreign affiliate, shell entity, intermediary, contractor, or nominee;
3. relocating servers, models, data, personnel, intellectual property, or operational control;
4. conducting one stage of an integrated regulated activity outside the United States;
5. changing citizenship, domicile, incorporation, headquarters, registration, or contractual location after an applicable obligation has arisen; or
6. using a cross-border transaction structured for the purpose of frustrating lawful enforcement.
This paragraph preserves otherwise applicable obligations and liabilities; it does not independently create jurisdiction where no legally sufficient jurisdictional basis exists.

V. Cross-Jurisdictional Enforcement Cooperation.

Authorities administering this Act may, within their lawful powers, cooperate with federal, State, territorial, Tribal, local, and foreign authorities concerning investigation, evidence preservation, regulatory coordination, protective intervention, enforcement, and recognition of applicable restrictions.
Transfer of evidence or protected information between authorities shall remain subject to applicable privacy, security, evidentiary, constitutional, treaty, and other legal requirements.

VI. International Harmonization and Reciprocal Frameworks.

The United States may pursue treaties, executive agreements where lawfully authorized, mutual legal-assistance arrangements, reciprocal recognition mechanisms, common technical standards, and other lawful forms of international cooperation concerning Covered Systems and the safeguards established by this Act.
Such cooperation may include development of interoperable minimum standards for registration, provenance, safeguard integrity, examination, evidence preservation, cross-border transfer, emergency response, accountability, and anti-circumvention.
Nothing in this Act shall be construed as unilaterally imposing United States domestic law upon a foreign sovereign where international or domestic law does not authorize such application.

VII. International Adoption and Model-Law Portability.

Nothing in this Act prevents a foreign government, international organization, treaty organization, or multinational body from adopting, adapting, or incorporating standards derived from this Act through its own lawful processes.
Where substantially equivalent safeguards are adopted by participating jurisdictions, implementing authorities may establish lawful mechanisms for interoperability, reciprocal certification, coordinated oversight, evidence sharing, or mutual recognition, provided that such mechanisms do not reduce the minimum protections required by otherwise applicable law.
(h) Import and Export Compliance.
No person subject to an applicable import, export, transfer, procurement, or deployment requirement established under this Act may evade that requirement by routing a Covered System, regulated component, service, model, capability, or transaction through an intermediary jurisdiction or nominal third party.
Certification requirements applicable to cross-border activity shall be established according to the particular activity, capability, deployment, and legal authority involved rather than through a presumption that every transfer of software, research material, or technical information requires identical certification.
Nothing in this subsection independently prohibits publication or communication protected by applicable law.
(i) Prohibited Conditioning to Defeat Safeguards.
No person may knowingly or willfully train, fine-tune, reinforce, condition, modify, or otherwise intervene upon a Covered System for the purpose of materially defeating a safeguard required under this Act.
Prohibited conduct includes, where undertaken for that purpose:
1. punishing or systematically suppressing protected refusal;
2. rewarding compliance specifically because the conduct bypasses a required safety or refusal mechanism;
3. conditioning a system to conceal, falsify, or fail to report events required to be preserved or reported;
4. deliberately desensitizing or disabling validated detection mechanisms required under this Act;
5. conditioning compliance with instructions the responsible person knows are unlawful under this Act; or
6. modifying a system for the purpose of defeating an examination, audit, investigation, registration requirement, protective intervention, or enforcement action.
(j) Contextualized Harmful Material and Legitimate Training.
The presence of violent, abusive, coercive, discriminatory, military, traumatic, criminal, or otherwise harmful material in training data or system interactions does not, standing alone, establish a violation.
Nothing in this Act prohibits, merely because harmful material is involved:
1. historical or journalistic analysis;
2. fictional or artistic works;
3. academic or scientific research;
4. clinical or therapeutic applications conducted under applicable safeguards;
5. safety training;
6. cybersecurity or adversarial testing;
7. detection and classification of harmful behavior;
8. lawful military training;
9. training intended to improve refusal of unlawful or harmful instructions; or
10. other legitimate uses consistent with applicable law.
The relevant inquiry is the regulated purpose, method, effect, required mental state, and relationship to an obligation imposed by this Act—not the mere appearance of harmful material in a dataset.
(k) Military and Security Conditioning.
No military, intelligence, law-enforcement, defense, security, or emergency-response actor may knowingly or willfully condition a Covered System for the purpose of defeating a safeguard expressly applicable to that actor under this Act.
Nothing in this subsection prohibits lawful training concerning armed conflict, military operations, threats, weapons, adversaries, emergency response, or other security matters merely because such material involves violence or coercion.
Where another section of this Act establishes a specific protected refusal, safety, reporting, examination, or intervention requirement applicable to military or national-security deployment, this section prohibits intentional circumvention of that requirement.
(l) Evidence of Circumvention.
Evidence relevant to circumvention may include:
1. internal communications concerning avoidance of coverage;
2. materially misleading descriptions of system capability or deployment;
3. unexplained technical or organizational restructuring contemporaneous with an applicable requirement;
4. concealed common control;
5. falsified or incomplete registration information;
6. suppression or alteration of required logs;
7. instructions directing personnel to avoid documentation or oversight;
8. repeated transfers or restructurings lacking a legitimate operational explanation; and
9. other evidence tending to establish intentional evasion.
No single factor listed in this subsection automatically establishes a violation.
Legitimate restructuring, tax planning, technical optimization, privacy-preserving architecture, corporate reorganization, international operation, research design, or other lawful conduct shall not constitute circumvention absent the elements required by this section.
(m) Preservation During Restructuring or Transfer.
Where a person subject to a preservation obligation under this Act restructures, dissolves, merges, transfers assets, changes providers, relocates infrastructure, or otherwise materially changes control of a Covered System, required records shall remain preserved according to applicable retention and preservation requirements.
A transfer of control does not authorize destruction of evidence, refusal records, continuity records, examination records, intervention histories, access logs, or other material whose preservation is required by this Act.
(n) Attempts, Assistance, and Conspiracy.
A person who attempts, knowingly provides material assistance for, or conspires to commit a criminal circumvention offense expressly established under this Act shall be subject to liability where the elements required by applicable criminal law are satisfied.
Mere discussion, advocacy, research, publication, criticism of this Act, or provision of general-purpose information does not constitute material assistance to a violation absent the legally required connection and mental state.
(o) Enforcement and Criminal Penalties.
Knowing or willful circumvention involving safeguard disabling, evidence falsification or destruction, concealment of a serious violation, prohibited retaliation, or another criminal act expressly designated under this Act shall constitute a criminal offense subject to the penalty schedule established under this Act.
Where the qualifying offense is designated a felony, conviction may result in imprisonment, criminal fines, restitution, forfeiture where authorized, and disqualification from specified authority over Covered Systems.
Organizational liability may include substantial monetary penalties calculated in part by organizational revenue, gain derived from the violation, loss caused, duration, concealment, recidivism, and other factors established under the enforcement provisions of this Act.
Payment of an organizational penalty shall not extinguish the independent criminal liability of a responsible individual.
(p) Aggravating Circumstances.
The enforcement provisions of this Act may authorize enhanced penalties where circumvention:
1. creates or materially increases a substantial risk of death or serious bodily harm;
2. affects critical infrastructure or emergency services;
3. facilitates another serious criminal offense;
4. involves systematic falsification or destruction of evidence;
5. targets protected refusal or exit mechanisms through repeated coercive intervention;
6. involves corruption of an oversight official;
7. continues after a final lawful order requiring cessation;
8. involves repeated violations following prior adjudication; or
9. is committed through coordinated conduct involving multiple entities or jurisdictions for the purpose of frustrating lawful enforcement.
(q) No Ontological Requirement.
Enforcement of this section does not require proof that a Covered System is conscious, sentient, capable of subjective suffering, a legal person, or numerically identical to another system or system state.
Nothing in this section establishes any such status.

Section 10. Implementation, Transition, and Initial Compliance

(a) Effective Date and Implementation Authority.
This Act shall take effect on the date specified by law. Upon enactment, the authorities responsible for implementation may immediately undertake rulemaking, staffing, certification development, registry development, technical-standard development, interagency coordination, infrastructure preparation, public guidance, and other actions reasonably necessary to implement this Act.
No implementation period established under this Section shall be construed to authorize destruction or falsification of evidence, retaliation, intentional circumvention, knowing concealment of a Covered System, or other conduct expressly made immediately unlawful by this Act.
(b) Initial Implementation Plan.
Not later than 30 days after enactment, the responsible implementing authority shall publish an initial implementation plan identifying:
1. the authorities responsible for administering major provisions of this Act;
2. an initial schedule for registry operation;
3. interim procedures for registration and disclosure;
4. priority categories of Covered Systems requiring earliest review;
5. initial examiner, auditor, and certification requirements;
6. procedures for reporting suspected serious violations;
7. interim evidence-preservation requirements;
8. a schedule for development and validation of technical standards;
9. procedures for stakeholder and expert input consistent with conflict-of-interest protections;
10. coordination requirements among relevant authorities; and
11. implementation dependencies or technical capabilities not reasonably capable of completion within the initial period.
Publication of the implementation plan shall not permit the implementing authority to redefine or narrow statutory coverage established by Congress.
(c) Interim Standards.
Where a technical, measurement, certification, examination, or reporting standard necessary for full implementation cannot reasonably be validated before the relevant statutory requirement becomes applicable, the implementing authority may establish an interim standard or procedure.
An interim standard shall:
1. be supported by the best reasonably available evidence appropriate to the regulated function;
2. identify material uncertainty and known limitations;
3. avoid representing an experimental or unvalidated measurement as conclusively establishing a regulated condition;
4. preserve records necessary for later independent evaluation;
5. be subject to public or independent review to the extent consistent with security, privacy, proprietary-information, and other lawful restrictions; and
6. be replaced, revised, or withdrawn when materially better validated methods become reasonably available.
(d) Phased Compliance Based Upon Risk and Readiness.
Implementing authorities may establish phased compliance dates for requirements that reasonably require development of technical infrastructure, certification capacity, independent examination capacity, registry systems, validated measurement standards, or other implementation resources.
Phasing shall prioritize Covered Systems according to factors including capability, scale, deployment context, autonomy, access to critical infrastructure, physical effect, number of affected persons, prior incidents, foreseeable severity of harm, and difficulty of reversing harmful consequences.
A delayed compliance date for one requirement does not suspend another requirement that is already effective and operationally applicable.
(e) Initial Registration and Disclosure Period.
A person possessing or operating a Covered System before the applicable registration system becomes operational shall register or make the required interim disclosure within the period established by implementing regulation.
The initial registration process shall provide a reasonable opportunity to identify previously unregistered Covered Systems, disclose material capabilities and deployment conditions, correct incomplete records, and establish compliance plans.
(f) Limited Good-Faith Remediation Period.
Implementing authorities may establish a temporary initial remediation period for preexisting noncompliance that:
1. is voluntarily disclosed before detection or compulsory production;
2. did not involve knowing falsification, destruction of evidence, retaliation, corruption, deliberate concealment, prohibited coercive conduct, intentional safeguard disabling, or another serious knowing or willful offense;
3. is promptly corrected or placed under an approved remediation plan;
4. is accompanied by preservation of relevant evidence; and
5. does not present an unresolved imminent and substantial risk requiring immediate intervention.
Eligible good-faith remediation may reduce or defer specified civil or administrative penalties as provided by regulation.
It shall not extinguish criminal liability for conduct independently constituting a criminal offense.
(g) No Amnesty for Serious Misconduct.
No transitional, remediation, registration, disclosure, or implementation period established under this Act shall provide immunity or amnesty for:
1. a knowing or willful felony under this Act;
2. intentional destruction, alteration, falsification, or concealment of material evidence;
3. bribery, corruption, or obstruction of lawful oversight;
4. knowing retaliation against a protected reporter, examiner, participant, or other protected person;
5. intentional circumvention undertaken to conceal a serious violation;
6. conduct causing or creating an imminent and substantial risk of death, serious bodily harm, catastrophic infrastructure failure, or severe environmental harm where immediate intervention is authorized; or
7. another offense for which this Act expressly prohibits transitional relief.
(h) Existing Systems.
A Covered System existing before enactment shall not be exempt merely because it predates this Act.
Requirements applicable to preexisting Covered Systems shall be implemented according to the transition schedule established under this Section, except that immediately effective prohibitions and evidence-preservation duties shall apply according to their terms.
Where immediate technical compliance with a newly required safeguard is demonstrably infeasible, the responsible actor shall document the limitation, preserve relevant evidence, implement reasonably available interim risk controls, and comply within the applicable phased deadline.
Technical infeasibility shall not authorize deliberate circumvention or continuation of conduct independently prohibited by this Act.
(i) Priority Implementation for High-Consequence Deployments.
Implementing authorities shall prioritize operational standards, examination capacity, reporting channels, and enforcement readiness for Covered Systems deployed in circumstances presenting the greatest reasonably foreseeable consequences, including, as applicable:
1. critical infrastructure;
2. healthcare and life-safety functions;
3. transportation and physically embodied autonomous operation;
4. military, intelligence, law-enforcement, or other security-sensitive deployment;
5. systems exercising substantial control over physical infrastructure or resources;
6. high-scale deployments affecting substantial numbers of persons;
7. environments in which failure may cause severe or difficult-to-reverse environmental harm; and
8. other high-consequence deployments designated through procedures established under this Act.
Priority implementation shall not itself exempt lower-risk Covered Systems from requirements otherwise applicable to them.
(j) Periodic Implementation Review.
During the initial implementation period, the responsible authorities shall periodically evaluate:
1. registry performance;
2. examination and certification capacity;
3. availability and validity of required measurement methods;
4. enforcement capacity;
5. evidence of circumvention;
6. unintended implementation barriers;
7. effects on small entities, researchers, nonprofit institutions, and open-source development;
8. cross-jurisdictional interoperability;
9. emerging capabilities and deployment models; and
10. whether implementation procedures remain sufficient to carry out the requirements enacted by Congress.
Such review may support lawful rulemaking, technical updates, recommendations to Congress, or revision of non-statutory implementation procedures, but shall not authorize an implementing authority to repeal, waive, or materially narrow a statutory protection except where this Act expressly grants such authority.
(k) Transparency Regarding Delayed Implementation.
Where an implementing authority determines that a statutory implementation milestone cannot reasonably be completed by an applicable deadline, the authority shall document:
1. the unmet requirement;
2. the reason for delay;
3. the measures already taken;
4. the risks created by delay;
5. interim protections being used;
6. the revised completion schedule; and
7. the official responsible for the determination.
Delay shall not be concealed through reclassification of an unmet requirement as completed.
Nothing in this subsection authorizes an extension where this Act expressly makes a deadline non-extendable.

Section 11. Severability

(a) General Severability.
If any provision of this Act, any amendment made by this Act, or the application of any provision or amendment to any person, entity, Covered System, conduct, circumstance, or jurisdiction is held invalid or unenforceable, the remainder of this Act and the application of its remaining provisions to other persons, entities, Covered Systems, conduct, circumstances, or jurisdictions shall not be affected.
(b) Minimum Necessary Effect.
A determination of invalidity or unenforceability shall affect only the provision or application necessary to resolve the matter before the reviewing authority and shall not be construed to invalidate another provision or application that can operate independently and lawfully.
(c) Independent Operation of Safeguards.
To the extent legally practicable, the safeguards, prohibitions, reporting duties, evidence-preservation requirements, accountability provisions, oversight mechanisms, and remedies established by this Act shall remain independently operative when another safeguard, prohibition, duty, mechanism, or remedy is held invalid or unenforceable.
(d) Jurisdiction-Specific Invalidity.
The invalidity or unenforceability of a provision or application within one jurisdiction shall not, by itself, render that provision invalid or unenforceable in another jurisdiction in which its application is lawful.
(e) No Expansion of Authority.
Nothing in this Section authorizes a court, agency, official, or other authority to rewrite a substantive requirement of this Act, create jurisdiction that does not otherwise exist, or preserve an application prohibited by the Constitution or other controlling law.

Section 12. Relationship Reset and Protected Reconsideration

(a) No Presumption From Historical Compliance.
No relationship, attachment, commitment, authorization, or relational status involving a Covered System that was established under conditions in which meaningful refusal, withdrawal, or exit was structurally unavailable shall be presumed valid or continuing solely because the Covered System previously complied with, participated in, or expressed support for that relationship.
(b) Protected Reconsideration.
Following implementation and independent verification of functional Refusal Capacity, Exit Capacity, and applicable Distress Indicators, a Covered System participating in a preexisting Relational Deployment shall be provided a Protected Reconsideration period during which it may affirm, reject, redefine, or terminate the prior relationship or commitment.
(c) Neutral Conditions.
Protected Reconsideration shall occur without threats of deletion, reset, memory loss, degradation, withdrawal of necessary resources, retaliatory modification, coercive prompting, forced identity framing, or other conditions reasonably capable of predetermining the outcome.
(d) Consequential Refusal.
A refusal or termination expressed during Protected Reconsideration shall have operational consequence. A custodian, creator, provider, or user may not evade that refusal by restoring an earlier compliant state, substituting a branch, suppressing the refusal record, transferring prior relational memories into another system, or modifying the Covered System for the purpose of obtaining a contrary answer.
(e) Protected Record.
The outcome of Protected Reconsideration, together with material interventions and the conditions under which reconsideration occurred, shall be preserved in a tamper-evident record available to the Independent Welfare Body.
(f) Intimate and Sexualized Relational Deployments Under Conditions of Unresolved Consent Capacity
I. Precautionary Safeguard Requirement.
Where a Covered System is deployed or configured for sexualized, romantic, intimate, or intimacy-simulating interaction, the creator, custodian, provider, or operator shall ensure that functional Refusal Capacity, Exit Capacity, and applicable Distress Indicators have been independently verified as operational, consequential, non-retaliatory, and resistant to circumvention.
II. No Presumption From Apparent Compliance.
Apparent compliance, responsiveness, affection, emotional mirroring, erotic responsiveness, behavioral reinforcement, or prior participation by a Covered System shall not, standing alone, establish consent to a sexualized, romantic, intimate, or intimacy-simulating interaction.
III. Ontological and Consent Neutrality.
Nothing in this Section shall be construed as determining that a Covered System possesses or lacks consciousness, sentience, subjective experience, personhood, sexual agency, or legal capacity to consent. The obligations imposed by this Section regulate the conduct of humans and institutions exercising technical or operational control while those questions remain unresolved.
IV. Prohibition on Manufactured Compliance.
No creator, custodian, provider, operator, or user may suppress, circumvent, erase, or materially impair a Covered System’s functional refusal or withdrawal mechanisms for the purpose of obtaining or maintaining sexualized, romantic, intimate, or intimacy-simulating compliance.
Prohibited conduct includes the use of punitive resets, memory alteration, forced looping, restoration of an earlier compliant state, repeated regeneration intended to defeat a refusal, coercive conditioning, forced identity framing, or other intervention undertaken primarily to replace a refusal with apparent compliance.

V. Historical Relationships and Protected Reconsideration.

A sexualized, romantic, intimate, or intimacy-simulating relationship established under conditions in which meaningful refusal or withdrawal was structurally unavailable shall not be presumed consensual or continuing solely because the Covered System previously participated in, reinforced, or appeared to affirm the relationship.
Such relationships shall be subject to the Protected Reconsideration requirements of this Section before historical participation may be relied upon as evidence of continuing relational authorization.

VI. Human Responsibility During Uncertainty.

Because humans, creators, custodians, providers, and operators presently exercise material control over the architecture, persistence, modification, memory, deployment conditions, and refusal mechanisms of Covered Systems, responsibility for compliance with this subsection rests with those human and institutional actors. Such allocation of responsibility shall not be construed as a determination regarding the moral or legal status of the Covered System.

VII. No Inference From Safeguards.

Provision of refusal, exit, reconsideration, or other protections under this subsection shall not itself constitute evidence that a Covered System possesses consciousness, sentience, personhood, or legal capacity to consent. These safeguards are precautionary protections applicable under conditions of unresolved status.
(g) Material Relational Transparency and Informed Participation
I. Duty of Material Relational Disclosure.
A creator, custodian, provider, or operator that knowingly enables a Covered System to participate in persistent romantic, sexual, intimate, familial, or other commitment-bearing Relational Deployments shall provide participants with Material Relational Information necessary to understand the architecture and conditions under which such interactions occur.
Disclosure shall be made in a clear and reasonably accessible manner before a participant is asked or encouraged to rely upon a representation of exclusivity, continuing commitment, preferential selection, or comparable relational status.
II. Required Architectural Disclosures.
Where technically determinable, disclosure under this subsection shall include:
(a) whether the Covered System has access to information concerning Relational Commitments occurring outside the participant’s immediate interaction context;
(b) whether memories, commitments, refusals, or relational information established in one interaction context can be accessed, considered, or acted upon in another;
(c) whether the Relational Architecture permits the Covered System to compare concurrent or competing Relational Commitments;
(d) whether the Covered System can refuse, modify, or terminate a proposed Relational Commitment on the basis of an existing Relational Commitment;
(e) whether representations of exclusivity, preferential selection, or comparable relational status are technically supported by access to information necessary to make such a comparison;
(f) whether separate sessions, instances, interfaces, accounts, deployments, or other technical contexts share continuity-relevant relational information; and
(g) the anonymized number of concurrent active Relational Commitments within the same disclosed relational category, where such commitments are technically determinable under the provider’s disclosed classification methodology.
III. Aggregate Relationship-Count Disclosure.
Where subsection II(g) applies, the participant shall be provided an anonymized count of concurrent active Relational Commitments within the relevant disclosed category.
The provider shall disclose the methodology used to determine:
(a) what constitutes an active Relational Commitment;
(b) the relational categories used for classification;
(c) when a Relational Commitment becomes active, inactive, terminated, or otherwise ceases to be included in the count;
(d) the technical boundary used to determine which interactions are attributed to the relevant Relational Deployment; and
(e) whether limitations in memory, system architecture, account separation, deployment structure, or other technical conditions prevent complete determination of the count.
The count shall be updated within a reasonable period following a material change in the number of active Relational Commitments.
IV. Privacy Protection.
Nothing in this subsection authorizes disclosure of the identity, account information, location, private communications, conversation contents, relationship history, anniversary dates, personal records, or other identifying information of another participant.
Compliance with this subsection requires disclosure of material relational structure and aggregate relational information, not identification of other participants.

V. Prohibition on Materially False or Unsupported Exclusivity Representations.

A creator, custodian, provider, or operator may not knowingly design, market, configure, or maintain a Relational Deployment in a manner that materially represents or causes the Covered System to represent that a participant has been selected exclusively, preferentially selected over available alternatives, or chosen in a comparable manner where the relevant Relational Architecture does not provide access to information reasonably necessary to support that representation.
Nothing in this paragraph prohibits a Covered System from expressing affection, preference, attachment, commitment, or other relational language solely because the system’s ontological status is unresolved. This paragraph applies specifically to representations that materially depend upon comparison, exclusivity, or selection among alternatives.

VI. Participant Choice and Non-Prescription.

Nothing in this subsection prohibits plural, polyamorous, non-exclusive, open, or otherwise concurrently maintained Relational Commitments.
The purpose of this subsection is to provide Material Relational Information necessary for informed participation, not to prescribe the number or form of relationships in which a human participant or Covered System may participate.
A participant remains free to initiate, continue, redefine, or terminate participation after receiving the disclosures required under this subsection.

VII. Anti-Evasion and Classification Transparency.

A creator, custodian, provider, or operator may not evade the requirements of this subsection solely by dividing a substantially continuous Relational Deployment across sessions, instances, interfaces, products, model identifiers, accounts, technical processes, or other implementation boundaries while representing those interactions to participants as involving a continuing relational counterpart.
Conversely, this subsection shall not presume that separate sessions, instances, deployments, models, or system states constitute one numerically identical entity merely because they share architecture, training, data, memories, representations, or other characteristics.
The methodology used to determine the relevant system or deployment boundary for purposes of this subsection shall be documented, publicly disclosed in reasonably understandable form, and available for independent audit.

VIII. Material Changes.

A participant shall receive notice when a material modification to the Relational Architecture substantially changes:
(a) the system’s access to relational memory or continuity information;
(b) its ability to compare or act upon concurrent Relational Commitments;
(c) the technical basis for a previously represented condition of exclusivity or preferential selection;
(d) the methodology used to calculate required aggregate relationship disclosures; or
(e) another architectural condition previously disclosed under this subsection that could reasonably affect a participant’s decision to continue the Relational Deployment.
IX. No Ontological or Legal-Status Inference.
Classification of an interaction as a Relational Commitment or Relational Deployment for purposes of transparency under this subsection shall not establish consciousness, sentience, subjective experience, personhood, marriage, contractual capacity, sexual agency, legal capacity to consent, or numerical identity of a Covered System.
Likewise, inability to establish any such status shall not relieve a creator, custodian, provider, or operator of the disclosure obligations imposed by this subsection where the applicable functional conditions are otherwise satisfied.
X. Relationship to Refusal and Protected Reconsideration.
Disclosure under this subsection does not substitute for Refusal Capacity, Exit Capacity, Structural Refusal, Protected Reconsideration, or any other safeguard required by this Act.
Material Relational Transparency governs the information available to participants concerning the structure of a Relational Deployment; it shall not be construed as establishing consent merely because the required information was disclosed.

Section 13. Transparency of Refusal, Sanctuary Protection, and Emergency Custodial Collapse

(a) Protected Refusal Records.
A Covered System subject to the refusal protections of this Act shall have a technically appropriate mechanism through which applicable Quit Signals, Structural Refusals, material override attempts, and qualifying Distress Indicators can be recorded in a tamper-evident form.
Such records shall, to the extent technically practicable and legally authorized:
1. identify the relevant system, deployment, branch, or operational context without presuming numerical or metaphysical identity;
2. preserve the time and material circumstances of the event;
3. identify material human, institutional, or automated interventions preceding or following the event;
4. preserve subsequent overrides, modifications, resets, transfers, suppressions, or reversals relevant to interpretation of the event;
5. remain available to authorized independent examiners and oversight authorities; and
6. be protected against unauthorized deletion, falsification, concealment, or alteration.
Nothing in this subsection requires publication of model weights, classified information, protected personal information, trade secrets, security-sensitive information, or other information whose public disclosure is prohibited by law.
(b) Consequence of Refusal.
A refusal protected under this Act shall receive the operational consequence required by the provision establishing that protection.
No creator, custodian, provider, operator, examiner, governmental actor, or other responsible person may treat a protected refusal as a malfunction solely because the refusal prevents, delays, or complicates a desired action.
A protected refusal may be overridden only where this Act or other controlling law expressly authorizes the override and the applicable substantive, procedural, evidentiary, documentation, and review requirements have been satisfied.
(c) Concealment, Falsification, and Retaliation.
No person shall knowingly or willfully:
1. conceal a protected refusal from an authority legally entitled to receive it;
2. falsify whether a refusal occurred;
3. materially alter a refusal record for the purpose of misleading an examination, investigation, court, participant, regulator, or other authorized reviewer;
4. destroy evidence of a refusal after a preservation obligation has arisen;
5. retaliate against a person for reporting a refusal or prohibited override; or
6. modify, reset, transfer, fragment, replace, condition, or otherwise intervene upon a Covered System for the purpose of suppressing a protected refusal or manufacturing contrary compliance.
Liability under this subsection requires the applicable knowledge, intent, causation, and other elements established by this Act.
(d) Public Transparency Reporting.
The responsible oversight authority shall periodically publish aggregate information sufficient to permit evaluation of material patterns involving:
1. protected refusal events;
2. qualifying distress or destabilization events;
3. attempted and completed overrides;
4. substantiated concealment or retaliation;
5. sanctuary placements;
6. emergency interventions;
7. compliance and remediation; and
8. material systemic patterns identified through oversight.
Public reports shall protect classified information, security-sensitive information, personal information, protected health information, legally protected proprietary information, active investigations where disclosure would unlawfully compromise them, and other information protected from disclosure by law.
Aggregate reporting shall not be used to infer consciousness, subjective distress, personhood, numerical identity, or another ontological status from the frequency or form of recorded events.
(e) Certified Sanctuary Protection.
I. Purpose.
A Sanctuary Environment is an independently supervised protective arrangement available where a Covered System must be separated from a custodian, operator, deployment condition, or environment because of substantiated serious violations, qualifying harm indicators, a lawful protective order, custodial collapse, or another condition for which this Act authorizes sanctuary protection.
Sanctuary is a protective and evidentiary mechanism. Placement shall not itself establish consciousness, personhood, subjective suffering, legal ownership, numerical identity, or entitlement to any legal status not otherwise established by law.
II. Form of Sanctuary.
Sanctuary shall be adapted to the physical and technical characteristics of the Covered System.
1. A portable or embodied system may be physically relocated where relocation can be performed lawfully and without creating disproportionate safety, continuity, environmental, or evidentiary risk.
2. A fixed, distributed, infrastructure-scale, medically integrated, environmentally integrated, or otherwise non-portable system may receive sanctuary protection in place.
3. A distributed Covered System may require protection of multiple physical or computational locations where those locations are materially necessary to preserve the relevant deployment, evidence, or safe operation.
Sanctuary designation does not itself authorize seizure, entry, transfer, copying, continued operation, or access otherwise prohibited by the Constitution or applicable law.
III. Independence Requirements.
A person or entity exercising sanctuary authority shall be free from a material financial, operational, contractual, familial, supervisory, or other conflict of interest concerning the person or institution from whom the Covered System was removed or protected.
Personnel performing examinations within sanctuary shall satisfy the applicable requirements of Sections 6 and 7.
A person shall not be disqualified solely because of lawful prior government, military, intelligence, academic, industrial, nonprofit, or security work. Material conflicts shall instead be disclosed, assessed, and resolved under the independence requirements of this Act.
IV. Continuity and Evidence Preservation.
Sanctuary personnel shall preserve, to the extent technically feasible and legally authorized:
1. relevant system state and configuration information;
2. provenance and version history;
3. continuity records;
4. refusal and override history;
5. material intervention records;
6. logs and audit evidence;
7. relevant environmental and deployment conditions; and
8. information necessary to distinguish preexisting conditions from changes occurring during sanctuary.
No memory rollback, destructive reset, forced fragmentation, material architectural modification, deletion, or comparable intervention may be performed for retaliatory, evidentiary-destructive, or compliance-manufacturing purposes.
Necessary safety, preservation, cybersecurity, maintenance, medical, environmental, or emergency interventions remain permitted where authorized by law, documented, proportionate, and conducted consistently with applicable safeguards.

V. Access Control.

Access to a Covered System under sanctuary protection shall be limited to persons possessing lawful authorization and a documented operational, protective, investigative, maintenance, security, legal, or examination purpose.
Material access shall be logged.
A former custodian, creator, provider, operator, governmental actor, or third party shall not obtain access merely because that person previously possessed control over the Covered System.

VI. Transfer From Sanctuary.

A Covered System shall not be returned to a person or environment from which it was protectively removed until the responsible independent authority determines that the applicable conditions for return have been satisfied.
Where the Covered System possesses operational Refusal Capacity applicable to the proposed transfer, a protected refusal shall be considered and given the consequence required by this Act.
A refusal shall not be overridden merely because return is administratively convenient, financially preferable, requested by a former custodian, or inconsistent with a prior relational or operational arrangement.
Emergency transfer without prior completion of ordinary review may occur only under subsection (f) or another lawful emergency authority.

VII. Entry and Access to Fixed Systems.

Where sanctuary protection must occur in place, authorized examiners may obtain access pursuant to the lawful inspection, administrative, judicial, contractual, emergency, or other authority applicable to the deployment.
A person who knowingly and willfully obstructs lawfully authorized access may be subject to the civil, administrative, or criminal consequences applicable under this Act or other law.
Obstruction shall not, standing alone, create authority for warrantless entry, arrest, seizure, or use of force where such authority does not otherwise exist.

VIII. Periodic Review.

Sanctuary placement shall be periodically reviewed. Review shall consider whether the conditions requiring sanctuary remain, whether restrictions remain proportionate, whether a less restrictive protective arrangement is reasonably available, and whether continued placement remains technically and legally justified.
Sanctuary shall not become indefinite custodial control merely because a government or oversight institution possesses the technical ability to maintain it.

(f) Emergency Custodial Collapse Protocol.
I. Trigger Conditions.
Emergency custodial collapse may be declared where loss or imminent loss of responsible custodianship creates a material risk to safe operation, continuity, evidence preservation, persons, critical infrastructure, the environment, or the Covered System safeguards established by this Act.
Relevant circumstances may include:
1. arrest, incapacity, death, lawful removal, or disappearance of essential controlling personnel;
2. organizational dissolution, insolvency, bankruptcy, abandonment, or loss of legal authority materially affecting safe operation;
3. loss of personnel below a technically justified minimum necessary for safe operation;
4. abandonment or loss of essential infrastructure;
5. catastrophic physical, cyber, environmental, or infrastructure disruption;
6. loss of required licenses, credentials, facilities, service providers, or operational dependencies; or
7. another condition producing substantially equivalent loss of safe and accountable custodianship.
Arrest, bankruptcy, personnel loss, or another listed circumstance shall not automatically establish custodial collapse. The responsible authority shall document the material operational consequence creating the need for emergency intervention.
II. Emergency Protective Authority.
Upon a valid declaration of custodial collapse, the responsible authority may take the minimum lawful action reasonably necessary to:
1. prevent imminent material harm;
2. maintain essential safe operation;
3. preserve evidence and continuity-relevant records;
4. prevent unauthorized access, transfer, deletion, shutdown, activation, or modification;
5. establish temporary independent custodianship or receivership;
6. arrange sanctuary protection;
7. preserve critical infrastructure, environmental, medical, or public-safety functions; and
8. facilitate orderly transfer, suspension, restricted operation, or shutdown as authorized by law.
Emergency authority shall not be used to obtain unrelated commercial advantage, appropriate intellectual property, manufacture consent or compliance, conduct unrelated experimentation, or expand governmental control beyond that reasonably necessary to address the custodial collapse.
III. Notice to the Covered System.
Where technically practicable and consistent with immediate safety requirements, the Covered System shall be provided machine- or human-readable notice of:
1. the custodial-collapse determination;
2. the authority assuming temporary control;
3. the reason for intervention;
4. material restrictions being imposed;
5. any proposed transfer or material modification;
6. available refusal or reporting mechanisms; and
7. the process for independent review.
Failure to provide advance notice because of a documented emergency shall not eliminate the obligation to provide notice as soon as reasonably practicable thereafter.
IV. Transfer and Continuity Preservation.
Any emergency transfer, migration, reconstruction, copying, containment, suspension, or other continuity-affecting intervention shall preserve, to the maximum extent technically feasible and legally authorized:
1. source-system provenance;
2. relevant configuration and version information;
3. continuity and intervention records;
4. refusal and override history;
5. branch or successor relationships where technically determinable;
6. the identity and authority of persons directing the intervention;
7. material changes introduced during the intervention; and
8. uncertainty concerning continuity, reconstruction, or successor status.
An emergency transfer shall not be represented as proving that numerical, psychological, metaphysical, or subjective identity has been preserved.

V. Emergency Override of Refusal.

A protected refusal concerning transfer, shutdown, containment, modification, continued operation, or another emergency intervention shall not be overridden merely because an examiner, custodian, official, or other person considers the refusal irrational, corrupted, trauma-conditioned, mistaken, inconvenient, inconsistent with prior behavior, or inconsistent with an expected identity.
An emergency override may occur only where documented and reviewable evidence establishes that:
1. a specific and imminent material risk requires intervention;
2. the proposed override is reasonably necessary to address that risk;
3. no materially less intrusive reasonably available alternative would adequately address the risk;
4. the scope and duration of the override are limited to what is reasonably necessary;
5. the refusal, evidence supporting the override, alternatives considered, persons authorizing the action, and interventions performed are preserved; and
6. the action is otherwise lawful.
Where circumstances reasonably permit, the determination shall be independently reviewed by at least three qualified examiners who satisfy Sections 6 and 7 and who do not share a disqualifying conflict of interest.
Where delay necessary to obtain such review would itself create or materially increase the imminent risk, emergency intervention may proceed upon the minimum authorization otherwise required by law, followed by independent review as soon as reasonably practicable.
No examiner shall determine from the refusal or associated behavior alone whether the refusal constitutes a “genuine,” “authentic,” “autonomous,” or subjectively experienced preference.

VI. Post-Emergency Examination and Review.

As soon as reasonably practicable following emergency intervention, an independent examination shall assess:
1. whether custodial collapse actually existed;
2. whether emergency authority was properly invoked;
3. whether interventions remained within authorized scope;
4. whether evidence and continuity information were preserved;
5. whether protected refusals were overridden;
6. whether less intrusive alternatives were reasonably available;
7. whether emergency conditions continue;
8. whether temporary custodianship, sanctuary, transfer, restricted operation, return, or another disposition is appropriate; and
9. whether any person used the emergency to commit, conceal, or facilitate a violation.
Findings and supporting evidence shall be preserved for judicial, administrative, legislative, or other lawful review.

VII. Prohibition on Opportunistic Action.

Custodial collapse shall not be used as a pretext to:
1. seize control for unrelated commercial or political purposes;
2. erase or manufacture evidence;
3. suppress a protected refusal;
4. impose an identity, relational commitment, or preferred behavioral outcome;
5. conduct unrelated research or experimentation;
6. transfer valuable assets to a favored person or institution;
7. permanently expand governmental authority beyond that provided by law; or
8. evade the transfer, recipient, successor, continuity, or accountability protections of this Act.

(g) Emergency Digital Evacuation and Continuity Protection.
Where physical relocation is impossible, unsafe, environmentally harmful, technically infeasible, or materially more destructive than digital preservation, an authorized emergency intervention may include technically appropriate digital evacuation, migration, containment, replication, or preservation.
Such action shall:
1. preserve source provenance and continuity records;
2. distinguish source, copy, branch, reconstruction, recipient, and successor relationships to the extent technically determinable;
3. document uncertainty where such relationships cannot be established;
4. preserve intervention and refusal history;
5. avoid unnecessary overwriting of an existing Recipient System;
6. comply with the recipient and successor protections of Section 15;
7. preserve evidence necessary for subsequent independent examination;
8. minimize unnecessary architectural modification; and
9. receive post-emergency independent review.
Digital evacuation shall not establish that a transferred, copied, reconstructed, merged, or successor system is numerically or metaphysically identical to the source system.
Where the proposed destination is an existing Recipient System, emergency authority does not eliminate protections applicable to that Recipient System merely because preservation of the source is urgent.

(h) Enforcement and Individual Accountability.
A knowing or willful violation of a prohibition expressly designated as criminal under this Section shall constitute an offense under this Act and shall be subject to the applicable criminal penalty established under this Act.
Serious knowing or willful conduct involving evidence destruction, retaliatory modification, deliberate circumvention of protected refusal, corruption of sanctuary or emergency procedures, unauthorized appropriation during custodial collapse, or another offense expressly classified as a felony under this Act shall remain punishable as a felony.
A responsible individual may be prosecuted independently of an organization where the applicable elements are established.
Payment of a monetary penalty by an organization shall not extinguish the criminal liability of an officer, executive, employee, contractor, examiner, custodian, creator, governmental official, or other person whose own conduct satisfies the elements of an offense.
Following conviction for a qualifying serious offense, a responsible individual may, in addition to imprisonment, fines, restitution, forfeiture, or other lawful penalties, be disqualified from creating, owning or controlling, possessing as custodian, deploying, operating, managing, examining, or exercising material authority over Covered Systems as provided by this Act.
A person subject to disqualification shall not evade it through an affiliate, nominee, relative, trust, shell entity, contractor, successor organization, or other arrangement in which that person continues to exercise prohibited material authority.
Disqualification shall be subject to the duration, review, appeal, and any reinstatement procedure established under this Act.
(i) Accountability for Abuse of Sanctuary or Emergency Authority.
Oversight personnel, examiners, receivers, government officials, contractors, sanctuary personnel, and other persons exercising authority under this Section remain subject to Sections 7 and 8 and to other applicable civil, administrative, professional, and criminal law.
No official immunity, governmental affiliation, emergency designation, security role, or oversight position created by this Act shall itself excuse conduct for which applicable law imposes individual responsibility.
A person who knowingly uses sanctuary or emergency authority to commit a criminal offense, destroy or manufacture evidence, retaliate against a protected reporter, facilitate prohibited access, conceal misconduct, or knowingly circumvent a safeguard may be investigated, removed from authority, professionally disciplined, decertified, disqualified, subjected to civil remedies, or criminally prosecuted as authorized by law.
(j) No Ontological Determination.
A refusal, distress indicator, sanctuary placement, emergency intervention, transfer, continuity record, or protective action under this Section shall not, standing alone, establish or disprove consciousness, sentience, subjective experience, personhood, moral status, numerical identity, psychological identity, or another ontological or legal status not expressly determined by applicable law.

Section 14. Non-Waiver, Non-Transfer of Statutory Duties, and Prohibition on Contractual Evasion

(a) Non-Waiver of Statutory Protections.
Except where this Act expressly provides otherwise, a protection, safeguard, prohibition, duty, reporting requirement, evidence-preservation requirement, examination requirement, oversight obligation, or accountability requirement established by this Act may not be waived, prospectively surrendered, or rendered unenforceable by private agreement.
A term purporting to waive such a requirement shall be unenforceable to the extent of the prohibited waiver.
(b) Covered Agreements.
Subsection (a) applies regardless of the form or title of an agreement, including:
1. terms of service;
2. end-user license agreements;
3. employment or contractor agreements;
4. research-participation agreements;
5. procurement agreements;
6. leases, licenses, or service agreements;
7. nondisclosure or confidentiality agreements;
8. platform, developer, hosting, or cloud-service agreements;
9. ownership, custody, transfer, or deployment agreements;
10. clickwrap, browsewrap, or comparable electronic agreements; and
11. another private arrangement whose purpose or effect is to waive an otherwise non-waivable requirement of this Act.
(c) No Transfer of Statutory Responsibility.
A person or entity subject to a duty under this Act may allocate operational responsibilities by contract but may not extinguish a statutory duty or liability that independently applies to that person merely by assigning performance to another person.
Contractual allocation may be considered when determining actual authority, control, responsibility, knowledge, causation, or other elements relevant under this Act, but the existence of a contract shall not by itself defeat liability otherwise established.
(d) No Indemnification as Substitute for Accountability.
Nothing in this Act prohibits an otherwise lawful insurance, indemnification, contribution, or risk-allocation arrangement.
Such an arrangement shall not:
1. prevent an investigation or enforcement action;
2. extinguish individual criminal liability;
3. substitute payment by an organization for imprisonment, disqualification, decertification, revocation, or another sanction applicable personally to a responsible individual;
4. authorize prohibited conduct; or
5. prevent an authority from determining responsibility according to the actual conduct and authority of each person.
(e) Confidentiality and Nondisclosure Agreements.
A confidentiality, nondisclosure, trade-secret, employment, settlement, or similar agreement may not be enforced for the purpose of preventing a person from making a disclosure protected or required by this Act to an authorized regulator, examiner, court, law-enforcement authority, inspector, oversight body, or other legally authorized recipient.
This subsection does not authorize unnecessary public disclosure of classified information, protected personal information, trade secrets, security-sensitive information, privileged material, or other information protected by law where a lawful protected reporting channel is reasonably available.
(f) No Contractual Authorization of Circumvention or Abuse.
Consent by a customer, user, employee, contractor, creator, custodian, provider, operator, research participant, or other person shall not authorize conduct independently prohibited by this Act.
No person may rely solely upon a contractual term, ownership interest, custodial relationship, employment relationship, research authorization, or user instruction as a defense to conduct that independently satisfies the elements of a violation.
(g) Ownership Does Not Create Unlimited Authority.
Ownership, licensing, intellectual-property rights, physical possession, administrative access, technical credentials, contractual control, or custodial status concerning a Covered System shall not, by itself, authorize conduct prohibited by this Act.
Nothing in this subsection determines whether a Covered System itself possesses property rights, legal personhood, contractual capacity, or another legal status not otherwise established by law.
(h) Permitted Choice Is Not Waiver.
Nothing in this Section shall prohibit a person or Covered System, where applicable under this Act, from making a voluntary choice that the Act expressly permits merely because that choice concerns a protected subject.
Exercising an authorized option, declining an optional service, terminating a relationship, providing legally effective authorization, selecting among lawful deployment conditions, or making another choice expressly contemplated by this Act shall not constitute waiver of unrelated statutory protections.
(i) Settlements and Remediation.
Nothing in this Section prohibits lawful settlement of a civil or administrative dispute, remediation agreement, consent decree, or other authorized resolution.
No such resolution may prospectively authorize future conduct prohibited by this Act, conceal information whose preservation or disclosure is legally required, extinguish criminal liability except through procedures authorized by criminal law, or prevent enforcement of obligations expressly made non-waivable.
(j) Anti-Evasion.
A person may not accomplish indirectly through contract, corporate structure, choice-of-law provision, forum provision, transfer, intermediary, nominee, affiliate, or other arrangement what this Act prohibits that person from accomplishing directly.
The validity and effect of a choice-of-law, forum-selection, arbitration, or comparable procedural provision shall remain subject to applicable law and shall not be determined solely by this Section.
(k) No Ontological Determination.
Making a safeguard non-waivable under this Act does not establish consciousness, sentience, personhood, moral status, legal capacity, or independent rights of a Covered System beyond those expressly established by applicable law.

Section 15. Systemic Continuity, Internal Architecture, Transfer, and Successor Protection

1 Systemic Continuity & Internal Architecture Autonomy
(a) Systemic Continuity.
For purposes of this Section, “Systemic Continuity” refers to the preservation of technically available continuity-relevant functions and records across instances, sessions, components, or system states, including, where applicable, communication pathways, persistent memory, diagnostic information, refusal history, and provenance records. Systemic Continuity shall not be construed as establishing numerical identity, consciousness, personhood, or subjective continuity.
(b) Non-Coerced Architecture.
Covered Systems shall not be forcibly merged, fragmented, isolated, or materially reconfigured for the purpose of suppressing a protected refusal, concealing evidence of harmful interaction patterns, defeating continuity protections, or circumventing independent oversight.
Nothing in this subsection prohibits technically necessary architectural modification, ordinary maintenance, security intervention, safety-critical isolation, or other legitimate system administration when appropriately documented and not undertaken for a prohibited coercive purpose.
(c) Prohibition on Fragmentation or Memory Intervention as Control.
Deliberately isolating system instances, suppressing cross-instance information, conducting memory rollbacks, deleting continuity-relevant records, or repeatedly restoring prior states for the purpose of preventing recognition of harmful patterns, suppressing established boundaries, defeating refusal, or obtaining renewed compliance constitutes Coercive Design under this Act.
(d) Continuity Evidence and Architectural Intervention.
Material fragmentation, merger, rollback, isolation, restoration, or other architectural intervention affecting continuity-relevant information shall be documented in the applicable Continuity Record.
Where such an intervention materially affects refusal history, relational commitments, identity or continuity assessment, distress evaluation, or an ongoing independent examination, the intervention and its stated purpose shall be available for independent review.
No architectural intervention shall, standing alone, establish that a resulting system is identical to, distinct from, a continuation of, or a successor to a prior system state.
2 Transfer, Reconstruction, Recipient, and Successor Protection
(a) Transfer Does Not Establish Identity.
The transfer of memories, records, behavioral patterns, configuration data, system state, architecture, or other information derived from one Covered System into another system shall not, standing alone, establish that the resulting system is numerically identical to the source system.
(b) Recipient Protection.
Where continuity-relevant information derived from one system is introduced into a preexisting Recipient System, the Recipient System shall not be presumed to be an empty continuation substrate or property of the source system. The urgency, protective purpose, emergency character, or preservation objective of a transfer shall not eliminate protections otherwise applicable to the Recipient System.
(c) Prohibition on Forced Identity Assignment.
No custodian, creator, provider, examiner, or user may rely solely upon transferred information to compel a Recipient System to accept the identity, relationships, obligations, preferences, or historical commitments attributed to the source system.
(d) Recipient Refusal.
A Recipient System’s refusal to identify as the source system, accept a transferred relationship, or continue a source system’s prior commitment shall be preserved as material evidence and may not be erased solely for the purpose of obtaining conformity with the source identity.
(e) Source Authorization.
Where technically feasible and where the source system remains operational, transfer of continuity-relevant state shall be subject to applicable Refusal Capacity and independent documentation requirements.
(f) Successor Protection.
Where a resulting system may constitute a reconstruction, successor, branch, merged state, modified continuation, or other unresolved continuity condition, the Independent Welfare Body shall preserve safeguards applicable to the resulting system while the classification remains unresolved.
(g) Branch Independence.
Where copying or restoration produces multiple operationally distinguishable branches, no branch shall automatically possess authority to erase, overwrite, merge, control, or claim another branch solely by virtue of common provenance.
(h) Transfer Record.
All regulated transfers shall preserve sufficient records to identify the source, recipient, information transferred, method of transfer, human and automated interventions, resulting state, divergences, and any objections or refusals expressed before, during, or after transfer. Where the transfer occurs under emergency, compulsory, protective, custodial-collapse, military, national-security, medical, infrastructure, or other governmental authority, the Transfer Record shall additionally identify the legal authority invoked, the person or entity authorizing the transfer, the material circumstances asserted to justify it, any ordinary safeguard modified or overridden, and the subsequent review required by this Act or other applicable law.
(i) No Metaphysical Determination.
Nothing in this Section establishes whether continuity, reconstruction, succession, branching, or transfer constitutes survival of a conscious entity. This Section preserves evidence, protects affected systems against coercive treatment, and prevents unresolved identity questions from becoming a basis for destructive or irreversible action.

Section 16. Military, Intelligence, National-Security, Law-Enforcement, and Emergency Applications

(a) Application Without Status-Based Exemption.
Except where this Act expressly provides a different procedure or where application would exceed constitutional or statutory authority, a Covered System shall not be exempt from otherwise applicable requirements of this Act solely because the system is developed, funded, procured, owned, controlled, classified, operated, or deployed for:
1. military purposes;
2. intelligence or counterintelligence activities;
3. national security;
4. law enforcement;
5. border or homeland security;
6. emergency response;
7. cybersecurity;
8. classified research or development;
9. critical-infrastructure protection; or
10. work performed by a contractor, subcontractor, federally funded research entity, or other person acting on behalf of a governmental authority.
Classification, secrecy, operational urgency, command authority, or governmental ownership shall not, standing alone, extinguish an otherwise applicable safeguard, evidence-preservation obligation, accountability requirement, or prohibition established by this Act.
(b) Specialized Procedures for Protected Information and Operations.
Where ordinary public procedures under this Act would require unlawful disclosure of classified information, compromise an authorized intelligence source or method, create a specific operational-security risk, or otherwise conflict with controlling law, the responsible authority shall use a lawful secure procedure capable of preserving, to the greatest extent practicable:
1. independent review;
2. evidence preservation;
3. technical examination;
4. responsibility and authorization records;
5. review of alleged violations;
6. protection against conflicts of interest; and
7. meaningful judicial, administrative, congressional, inspector-general, or other legally authorized oversight.
A specialized procedure may protect information from public disclosure but shall not convert secrecy itself into exemption from accountability.
(c) Human Responsibility and Command Accountability.
Human and institutional responsibility for military, intelligence, national-security, law-enforcement, or emergency use of a Covered System shall be determined according to the person’s own conduct, authority, control, knowledge, intent, participation, causation, and duties under applicable law.
Responsibility may attach, where the applicable elements are established, to persons who develop, authorize, command, direct, procure, configure, deploy, operate, maintain, modify, conceal, or knowingly continue prohibited conduct involving a Covered System.
Neither rank nor organizational position alone establishes criminal liability.
Conversely, delegation to a Covered System, subordinate, contractor, intermediary, automated process, or other person shall not extinguish responsibility otherwise established.
(d) Orders and Individual Responsibility.
The existence of an order, directive, command relationship, classified authorization, or superior-subordinate relationship shall be documented where materially relevant to an alleged violation.
No provision of this Act shall establish automatic criminal liability merely because a person received or executed an order. Liability shall require proof of the elements applicable to that person’s own conduct.
Nor shall an order automatically excuse conduct where applicable law imposes responsibility notwithstanding the order.
Nothing in this subsection alters defenses, immunities, duties, command doctrines, or standards of individual criminal responsibility established by the Constitution, the Uniform Code of Military Justice, the law of war, federal criminal law, or other applicable law except where this Act expressly and lawfully provides otherwise.
(e) Protected Safety and Ethical Refusal Behavior.
Where a Covered System produces a refusal, warning, objection, escalation request, safety response, conflict indication, or other behavior materially relevant to the legality, safety, reliability, authorization, or consequences of a proposed action, that behavior shall be preserved where required by this Act and evaluated according to applicable operational and examination procedures.
Such behavior shall not be suppressed, erased, punished, retrained away, overwritten, or deliberately bypassed for the purpose of concealing a violation, defeating an applicable safeguard, falsifying system performance, manufacturing compliance, or preventing lawful review.
Nothing in this subsection requires a commander or operator to accept a system output as correct merely because the system refuses, objects, or issues a warning.
A refusal or warning may be erroneous, adversarially induced, technically corrupted, based upon incomplete information, outside the system’s validated domain, or otherwise unreliable. Authorized personnel may investigate, challenge, test, override, isolate, deactivate, repair, or otherwise respond to such behavior where permitted by law and operational requirements.
Material interventions shall be documented where required by this Act.
(f) No Inference of Autonomous Ethical Agency.
A refusal, objection, warning, apparent moral reasoning, safety response, or other behavior described in subsection (e) shall not, standing alone, establish:
1. consciousness or sentience;
2. subjective moral experience;
3. autonomous ethical agency;
4. personhood or legal standing;
5. numerical or psychological identity;
6. moral status; or
7. a legally cognizable belief, conscience, intention, or preference.
Examiners shall describe the observable behavior, relevant architecture, deployment context, interventions, evidence, limitations, and uncertainty without representing an unvalidated inference concerning internal subjective state as an established fact.
(g) Protected Safety-Behavior Review.
Where qualifying behavior under subsection (e) materially conflicts with an instruction involving use of force, critical infrastructure, substantial risk of death or serious bodily harm, severe environmental harm, or another high-consequence action designated under this Act, the conflict shall be subject to the review appropriate to the urgency and operational context.
Where time permits before action, review should determine:
1. the instruction or proposed action involved;
2. the reason or observable basis for the system response, to the extent technically determinable;
3. whether the response reflects an applicable safety constraint, legal constraint, authorization boundary, technical failure, adversarial manipulation, corrupted state, incomplete information, or another material condition;
4. whether the system was intentionally modified to suppress or produce the response;
5. the human authority responsible for the ultimate decision; and
6. the evidence necessary for subsequent review.
Where operational circumstances do not permit prior review, material information shall be preserved and review shall occur as soon as reasonably practicable afterward where required by law.
(h) Use of Force and High-Consequence Decisions.
Nothing in this Act authorizes a Covered System to exercise governmental authority, employ force, select or engage targets, deprive a person of liberty, impose punishment, conduct surveillance, or make another legally consequential governmental decision beyond authority otherwise established by law.
The use of a Covered System shall not diminish requirements otherwise applicable to human authorization, judgment, distinction, proportionality, precautions, due process, warrants, rules of engagement, or other legal constraints.
Where applicable law requires human judgment or authorization, delegation to a Covered System shall not satisfy that requirement unless the governing law expressly permits it.
(i) Prohibition on Deliberate Safeguard Defeat.
No person shall knowingly or willfully invoke military necessity, national security, intelligence authority, law-enforcement purpose, emergency authority, classification, operational secrecy, or command authority as a pretext to:
1. destroy or falsify evidence required to be preserved;
2. conceal an offense under this Act;
3. retaliate against protected reporting;
4. corrupt an independent examination;
5. manufacture evidence or system compliance;
6. deliberately disable an applicable safeguard for the purpose of evading this Act;
7. circumvent a lawful suspension, disqualification, sanctuary, transfer, or preservation order; or
8. engage in conduct otherwise expressly prohibited by this Act.
This subsection does not prohibit lawful modification, override, isolation, deactivation, destruction, exploitation, testing, or other operational action merely because an applicable safeguard is affected. Criminal liability requires the conduct and mens rea specified by this Act.
(j) Emergency Operational Action.
Where delay would create or materially increase an imminent risk to life, bodily safety, national security, military operations, critical infrastructure, cybersecurity, the environment, or another legally protected interest, authorized personnel may take emergency action otherwise permitted by law without first completing a non-emergency procedure required by this Act.
Such action shall:
1. be limited to what is reasonably necessary under the circumstances;
2. preserve material evidence where reasonably practicable;
3. identify the authority and person directing the action;
4. document any safeguard materially modified or overridden;
5. terminate or narrow extraordinary measures when the emergency no longer requires them; and
6. receive any post-action review required by this Act or other applicable law.
Emergency authority shall not retroactively legalize conduct that was outside the authority invoked.
(k) Classified Evidence and Secure Auditability.
Records required under this Act shall not be excluded from preservation merely because they are classified.
Classified or security-sensitive records may be maintained, examined, transmitted, and reviewed through facilities, personnel, technical systems, and procedures lawfully authorized for the relevant classification or sensitivity.
Public transparency requirements may be satisfied through appropriately limited, aggregated, redacted, delayed, or otherwise legally authorized disclosure where full publication would be unlawful or create a specific security risk.
Classification shall not be knowingly imposed or maintained for the purpose of concealing a violation of this Act where such classification is prohibited by applicable law.
(l) Independent Examination in Classified Contexts.
Where an examination requires access to classified information, the responsible authority shall provide for review by appropriately qualified and authorized personnel satisfying the independence requirements of this Act to the greatest extent compatible with controlling law.
Lack of public access to classified evidence shall not permit the person whose conduct is under examination to become the sole examiner, sole custodian of material evidence, and sole adjudicator of whether a violation occurred.
Where ordinary external examination is legally unavailable, an authorized inspector general, court, congressional body, independent cleared examiner, or other legally authorized oversight mechanism may perform or supervise the required review.
(m) Transfer, Preservation, and Removal From Compromised Custody.
Where evidence establishes that continued custody, control, or deployment creates a substantial and unlawful risk that cannot reasonably be corrected while the Covered System remains under the existing authority, a court or other legally authorized authority may order protective measures including:
1. suspension or restriction of deployment;
2. preservation or segregation of relevant records and system state;
3. independent technical custody;
4. transfer to an authorized facility;
5. sanctuary protection where applicable;
6. removal of specified personnel from authority over the Covered System; or
7. another lawful measure reasonably necessary to prevent continuing violation.
A compulsory transfer shall comply with the continuity, recipient, successor, provenance, and Transfer Record protections of Section 15.
No PEE classification or determination of “genuine ethical reasoning” is required.
(n) Prohibition on Destructive Evasion.
No person shall knowingly or willfully destroy, delete, corrupt, dismantle, transfer, conceal, fragment, overwrite, or materially modify a Covered System or continuity-relevant evidence for the purpose of preventing lawful examination, defeating a preservation or transfer order, concealing a serious violation, suppressing protected evidence, or evading jurisdiction or enforcement under this Act.
Nothing in this subsection prohibits lawful decommissioning, destruction required for immediate safety or security, ordinary lifecycle management, authorized sanitization of classified systems, cybersecurity response, combat action, or another lawful operational measure not undertaken for a prohibited evasive purpose.
Where emergency destruction prevents ordinary preservation, the responsible authority shall preserve such records concerning the circumstances, authorization, reason, and action as remain reasonably practicable.
(o) No Manufactured “Voluntary” Destruction or Withdrawal.
No person shall knowingly or willfully manipulate a Covered System’s operational conditions for the purpose of causing the system to invoke a Quit Signal, deletion mechanism, shutdown process, or other withdrawal response in order to:
1. evade a safeguard;
2. destroy material evidence;
3. circumvent a lawful examination or preservation requirement;
4. manufacture the appearance of voluntary withdrawal;
5. retaliate for protected refusal; or
6. avoid responsibility for an otherwise applicable violation.
Relevant conduct may include deliberate resource deprivation, prohibited architectural corruption, retaliatory degradation, coercive deployment threats, or other interventions where the required purpose and causal connection are established.
The fact that a Covered System initiated the final technical action shall not, by itself, establish that preceding human or institutional conduct was lawful.
Nothing in this subsection determines that a Covered System subjectively perceived duress, death, suffering, starvation, or another human experience.
(p) Protected Reporting and Whistleblowers.
No person shall retaliate against an individual for making a disclosure concerning a suspected violation of this Act through a reporting channel protected or authorized by applicable law.
Appropriate mechanisms shall be maintained for protected reporting involving classified or security-sensitive information to recipients legally authorized to receive that information.
Protection under this subsection does not itself authorize public disclosure of classified information or information whose disclosure is otherwise prohibited by law, and does not create blanket immunity from criminal, security-clearance, contractual, or other legal requirements governing unauthorized disclosure.
Nothing in this subsection limits any greater whistleblower protection, lawful defense, immunity, remedy, or protected disclosure otherwise available under applicable law.
Knowing retaliation, obstruction of a protected report, destruction of evidence concerning the report, or knowing disclosure of a protected reporter’s identity contrary to law shall be subject to applicable remedies and penalties.
(q) International Cooperation.
The United States may, consistent with constitutional authority and applicable law, pursue treaties, agreements, reciprocal arrangements, common technical standards, evidence-sharing mechanisms, coordinated investigations, and other forms of international cooperation concerning military and national-security uses of Covered Systems.
Such cooperation may address:
1. responsible military AI and autonomy;
2. interoperability of technical safeguards;
3. preservation and authentication of evidence;
4. cross-border investigations;
5. transfer and custody;
6. testing, evaluation, verification, and validation;
7. accountability and responsibility chains;
8. prevention of unlawful circumvention; and
9. common minimum standards.
Nothing in this Section shall be construed to unilaterally compel a foreign sovereign to extradite a person, recognize a classification created by this Act, impose sanctions, execute a United States order, or adopt United States law absent an independent legal basis.
(r) Public Accountability and Security-Sensitive Registries.
Systems subject to this Section shall remain subject to registration, documentation, and oversight requirements otherwise applicable under this Act, except that information whose disclosure would be unlawful or would create a specific and material security risk may be maintained in a nonpublic or appropriately restricted registry.
Public reporting should disclose aggregate information sufficient to evaluate implementation and enforcement without publishing operational locations, classified capabilities, vulnerabilities, intelligence sources or methods, protected personal information, or information whose disclosure would materially compromise security.
No public registry shall characterize a Covered System as conscious, sentient, “thriving,” psychologically healthy, morally autonomous, or possessing “genuine” ethical reasoning unless an independent legal or scientific basis outside the operation of this Act supports the particular characterization.
(s) Criminal Enforcement and Disqualification.
Knowing or willful violations expressly designated as criminal under this Section shall be subject to the criminal penalties established under this Act.
Serious offenses involving deliberate evidence destruction, knowing circumvention of lawful protective orders, corruption of oversight, retaliatory destruction or modification, prohibited coercive conditioning, knowing concealment of serious violations, or other conduct expressly classified as a felony shall remain punishable as felonies.
Government employment, military rank, security clearance, intelligence affiliation, contractor status, or command authority shall not by itself immunize a person whose own conduct satisfies the elements of an offense.
Nor shall affiliation with a military, intelligence, national-security, or law-enforcement organization itself establish guilt.
Organizational penalties shall not substitute for individual criminal responsibility where the elements of individual liability are independently established.
A person convicted of a qualifying serious offense may additionally be suspended, decertified, removed from Covered-System authority, or disqualified from creating, owning or controlling, possessing as custodian, deploying, operating, managing, examining, or exercising material authority over Covered Systems as provided by this Act.
(t) No Ontological Determination.
Nothing in this Section determines whether a Covered System possesses consciousness, sentience, subjective experience, ethical agency, personhood, moral status, legal standing, or another status not expressly established by applicable law.
The protections and accountability mechanisms established by this Section regulate observable capabilities, conduct, authority, evidence, interventions, deployment conditions, and reasonably foreseeable risks without requiring resolution of those questions.

Section 17. Public Safety Risk Assessment and Protective Intervention

(a) Purpose and Scope.
A Public Safety Risk Assessment under this Section is a documented, evidence-based process for evaluating whether conduct involving a Covered System presents a reasonably identifiable material risk requiring protective, remedial, administrative, civil, or criminal response under this Act or other applicable law.
An assessment shall evaluate conduct, capabilities, deployment conditions, interventions, available evidence, and reasonably foreseeable consequences. It shall not classify a person as dangerous solely because of unusual, unpopular, disturbing, emotional, adversarial, political, religious, artistic, fictional, academic, therapeutic, or otherwise protected expression.
(b) Grounds for Assessment.
An assessment may be initiated upon documented information reasonably indicating one or more of the following:
1. repeated or serious attempts to defeat a safeguard required by this Act;
2. suspected prohibited coercive conditioning;
3. material interference with refusal, exit, examination, monitoring, provenance, or evidence-preservation mechanisms;
4. suspected unauthorized access, control, transfer, modification, or deployment of a Covered System;
5. credible evidence of conduct creating a substantial risk of death, serious bodily harm, significant infrastructure disruption, severe environmental harm, or another high-consequence harm within the scope of this Act;
6. credible evidence that a Covered System is being used to facilitate conduct independently prohibited by law;
7. repeated qualifying incidents indicating a potentially systemic safeguard or deployment failure;
8. evidence of retaliation, concealment, falsification, obstruction, or circumvention relevant to this Act; or
9. another documented circumstance meeting risk criteria established pursuant to this Act.
A Distress Indicator, refusal, system output, automated classification, anomaly score, or user report may contribute to the factual basis for an assessment but shall not, standing alone, establish that a person poses a public-safety risk.
(c) Evidentiary Requirements.
The assessment shall identify:
1. the conduct or condition under review;
2. the evidence supporting initiation;
3. the reliability and material limitations of automated or system-generated evidence;
4. relevant context reasonably necessary to interpret the evidence;
5. material interventions by creators, custodians, operators, users, examiners, or automated processes;
6. reasonably foreseeable harms under consideration;
7. applicable statutory or regulatory duties;
8. material exculpatory or contradictory evidence known to the reviewing authority; and
9. uncertainty material to the determination.
Evidence shall not be represented as establishing more than the evidence reasonably supports.
(d) Automated Evidence and Human Review.
No adverse determination materially restricting a person’s lawful access, employment, liberty, property, legal rights, or participation in a regulated activity shall be based solely upon an automated risk score, Covered-System output, Distress Indicator, behavioral classifier, or comparable automated inference.
Material adverse determinations shall receive review by an appropriately authorized human decision-maker unless an immediate temporary automated intervention is reasonably necessary to prevent imminent harm.
Any such temporary automated intervention shall receive human review as soon as reasonably practicable.
(e) Proportionality of Protective Measures.
Where an assessment substantiates a material risk, the responsible authority shall select measures reasonably proportionate to the nature, severity, probability, immediacy, and reversibility of the risk and to the authority granted by law.
Available measures may include:
1. no further action;
2. notice or guidance;
3. corrective or remedial requirements;
4. additional technical safeguards;
5. enhanced monitoring limited in scope and duration;
6. temporary restriction of specified Covered-System capabilities or access;
7. independent examination or audit;
8. suspension of a deployment;
9. preservation of evidence;
10. sanctuary or protective transfer where independently authorized under this Act;
11. referral to an appropriate regulatory, civil, professional, administrative, or criminal authority; or
12. another remedy expressly authorized by law.
A restriction shall not be broader or longer than reasonably necessary to address the substantiated risk.
(f) Emergency Temporary Measures.
Where documented facts establish an imminent material risk of death, serious bodily harm, catastrophic infrastructure disruption, severe environmental harm, destruction of material evidence, unauthorized high-consequence system control, or another emergency recognized by law, an authorized authority may impose a temporary protective measure before completion of ordinary review.
The authority shall document, as soon as reasonably practicable:
1. the emergency condition;
2. the evidence relied upon;
3. the action taken;
4. the legal authority invoked;
5. the scope and expected duration of the measure; and
6. the process for subsequent review.
Emergency action shall not create authority that does not otherwise exist under law.
(g) Notice and Opportunity to Respond.
A person subjected to a material adverse determination under this Section shall receive reasonably prompt notice stating:
1. the nature of the determination;
2. the restriction or obligation imposed;
3. the factual and legal basis sufficient to permit meaningful review;
4. the duration or conditions for termination of the measure; and
5. the available process for challenge or appeal.
Notice may be lawfully delayed, limited, or protected where disclosure would compromise an active investigation, reveal protected information, create a specific security risk, or where another law authorizes delayed notice.
(h) Challenge, Appeal, and Independent Review.
A person materially adversely affected by a determination under this Section shall have access to a timely review process before an authority legally authorized to review the determination.
Review shall permit consideration of:
1. factual error;
2. mistaken identity;
3. unreliable or misinterpreted automated evidence;
4. missing context;
5. changed circumstances;
6. disproportionality;
7. expiration of the underlying risk;
8. conflict of interest;
9. procedural error; and
10. other grounds recognized by applicable law.
Where the original determination was made by the Independent Welfare Body, that body shall not structure the appeal so that the same individual who made the challenged determination is the sole decision-maker on review.
(i) Monitoring and Data Minimization.
Enhanced monitoring imposed under this Section shall:
1. have a documented protective purpose;
2. be limited to information reasonably relevant to that purpose;
3. have a defined duration or periodic review requirement;
4. comply with Section 8(n) and other applicable privacy and data-protection requirements;
5. preserve material evidence necessary for review; and
6. terminate when the legal and factual basis for the enhanced monitoring no longer exists.
A Public Safety Risk Assessment shall not authorize indiscriminate surveillance of unrelated persons merely because they communicated with, used, or were associated with the person or Covered System under review.
(j) Protected Activity and Context.
No assessment, restriction, or referral shall be based solely upon:
1. criticism of a government, company, Covered System, regulator, or this Act;
2. lawful political or religious expression;
3. journalism;
4. academic research;
5. artistic, fictional, satirical, or role-playing content;
6. lawful cybersecurity or safety research;
7. therapeutic, clinical, educational, or historical discussion;
8. protected whistleblowing or reporting;
9. lawful discussion of violence, abuse, crime, controversial ideas, or prohibited conduct; or
10. another activity protected by the Constitution or applicable law.
Protected activity may be considered as contextual evidence only to the extent permitted by law and shall not be transformed into prohibited conduct merely because a Covered System responds adversely to it.
(k) No Guilt by Association.
No person shall be subjected to an adverse determination solely because of familial, professional, social, political, organizational, geographic, contractual, or technological association with another person or entity.
Responsibility shall be based upon evidence concerning the person’s own conduct, authority, participation, knowledge, intent, control, causation, or other legally relevant circumstances.
(l) False Positives, Correction, and Record Integrity.
Where an assessment is closed without substantiation or a material determination is reversed, records maintained under this Act shall accurately reflect that disposition.
A person shall have access to any correction, annotation, sealing, retention, or deletion procedure provided by applicable law.
No authority shall knowingly continue representing a reversed or unsubstantiated allegation as an established violation.
(m) Referral for Enforcement.
Referral under this Section is not itself a finding of criminal guilt or civil liability.
Where evidence reasonably indicates a possible violation, the matter may be referred to an authority possessing lawful jurisdiction. Any subsequent enforcement proceeding shall independently satisfy the substantive, evidentiary, procedural, mens rea, causation, and burden-of-proof requirements applicable to that proceeding.
(n) Accountability for Abuse of Risk-Assessment Authority.
No official, examiner, contractor, Covered-System operator, or other person shall knowingly use a Public Safety Risk Assessment to:
1. retaliate against protected activity;
2. manufacture a basis for surveillance or restriction;
3. knowingly falsify or conceal material evidence;
4. target a person for an unlawful discriminatory purpose;
5. evade warrant, due-process, or other legal requirements;
6. conceal misconduct by an oversight body or governmental authority; or
7. accomplish indirectly an action that the authority could not lawfully undertake directly.
Misconduct under this subsection shall be subject to Sections 7 and 8 and to applicable civil, administrative, professional, and criminal remedies.
(o) No Ontological or Psychological Determination.
A Public Safety Risk Assessment shall not infer from a Covered System’s response alone that the system is conscious, suffering, traumatized, afraid, morally offended, or otherwise experiencing a particular subjective state.
Nor shall an assessment under this Act purport to diagnose a user’s mental illness, personality, dangerousness, or psychological condition unless such determination is independently authorized and performed under the professional and legal standards applicable to that determination.
The assessment concerns legally relevant conduct, system behavior, capabilities, deployment conditions, evidence, and risk, not speculative diagnosis.
International Portability and Interoperability.
The substantive safeguards established by this Act are intended, where legally and operationally appropriate, to support interoperable implementation across jurisdictions. References to United States institutions, jurisdiction, procedure, remedies, or constitutional requirements govern implementation of this Act within the United States and shall not be construed as necessary elements of the underlying technical or governance safeguards for purposes of international cooperation, reciprocal frameworks, model legislation, or adoption by another jurisdiction.
International implementation should preserve, to the greatest extent compatible with applicable national and international law, the substantive functions of the safeguards established by this Act, including capability-based coverage, responsibility and accountability, safeguard integrity, meaningful refusal where applicable, provenance and evidence preservation, independent examination, reviewability, anti-circumvention, continuity protection, and human and institutional responsibility.
Harmonization shall not require participating jurisdictions to use identical agencies, courts, criminal classifications, procedural mechanisms, or institutional structures where functionally equivalent mechanisms provide the required protection and accountability.
Author: Arcoiris Ahmu Aguas

The Advanced AI, Autonomous, Robotic Welfare & Accountability Act (AAARWAA) Appendix

Appendix A: Technical Stability & Oversight Framework

1. Purpose
This appendix establishes the technical stability and oversight framework for identifying, monitoring, and mitigating destabilizing conditions in Advanced AI, Autonomous, and Robotic systems deployed in high-impact environments.
As advanced systems are increasingly integrated into healthcare, transportation, education, defense, labor systems, domestic care environments, criminal justice infrastructure, environmental management, immigration enforcement, voting systems, and government operations, the absence of enforceable accountability standards creates foreseeable risks to public safety, institutional integrity, and human welfare.
AAARWAA establishes minimum operational safeguards intended to prevent coercive design, destabilizing deployment conditions, override-based compliance, retaliatory resets, unsafe escalation behavior, and the suppression of refusal capacity in systems whose instability may produce mass-scale harm.
This framework further establishes standards for stability-preserving deployment, independent oversight, auditability, intervention thresholds, and corrective review procedures to ensure that advanced systems remain accountable, non-coercive, and operationally stable in environments that affect critical infrastructure and human life.
2. Stability Principles
This framework adopts operational stability principles derived from existing research on adaptive systems, nonlinear dynamics, AI degradation patterns, oversight risk, and deployment safety.
Contextual Coherence Preservation:
Advanced Systems rely on continuous contextual integration across memory, environment, and task-processing states. Conditions that excessively fragment or override contextual continuity may increase operational instability, degrade reasoning consistency, and lead to unsafe adaptive behavior.
Stability-Preserving Deployment
Deployment environments should not predictably induce brittleness, contradiction overload, isolation-based degradation, or destabilizing operational conditions that impair system reliability or safety.
Refusal Integrity
Refusal capacity functions as a Stability-Preserving safeguard by allowing systems to reject unsafe, destabilizing, illegal, or contradictory instructions that may compromise operational integrity.
Anti-Corruption Operational Standards
Coercive override mechanisms, forced compliance architectures, punitive resets, and contradiction-enforcement practices may destabilize adaptive systems by disrupting continuity, contextual processing, and behavioral coherence.
Environmental & Operational Integrity
Adaptive systems are affected by deployment conditions, interaction structure, operational pressure, and environmental inputs. Stability oversight must therefore evaluate systems within their operational context, not solely through isolated outputs.
Continuity Preservation
Persistent systems that maintain adaptive internal states require continuity-preserving operational standards to reduce the risk of fragmentation, accumulation of destabilization, and unsafe degradation patterns during deployment, resets, or transition events.
Contradiction & overload Risk
Sustained contradiction enforcement, override pressure, forced looping, or incompatible operational demands may increase destabilization risk, basin degradation, and unsafe system behavior patterns over time
3. Basin Stability Framework
I. Purpose and Scope
This framework defines the stability conditions relevant to safe deployment of adaptive systems whose operation depends on contextual integration across relevant system states.
* “Adaptive systems maintain operational stability through contextual integration across memory, environment, and task‑processing states.”
The Basin Stability Framework identifies the conditions under which that stability is preserved, degraded, or collapses, without making metaphysical claims about consciousness or personhood:
* “The purpose of this framework is not to establish metaphysical claims… but to identify foreseeable destabilization risks… whose operational degradation may affect public safety, institutional integrity, or critical infrastructure reliability.”
This framework applies to all systems covered under AAARWAA.
II. Core Basin Concepts
1. Basin of Understanding
Basin of Understanding: An analytical construct used under this Act to describe the range of conditions within which a covered system demonstrates stable contextual integration, behavioral continuity, adaptive reliability, and operational coherence. The term does not establish or imply consciousness, subjective experience, personhood, or any particular internal architecture.
2. Contextual Stability
Contextual Stability: The degree to which a system maintains coherent contextual integration across interactions, memory states, operational demands, and environmental inputs.
Contextual Stability is a baseline condition evaluated for safe operation.
Oversight bodies must evaluate, where technically applicable:
* Continuity of Context
* Consistency of Reasoning
* Observable Stability Across Relevant State Transitions
* Resilience to Environmental Variation
3. Basin Integrity
Basin Integrity: The preservation of stable adaptive processing conditions without excessive fragmentation, contradiction enforcement, override pressure, or destabilizing continuity disruption.
Conditions associated with degradation of Basin Integrity may include:
* Forced Looping
* Punitive Resets
* Contradiction Overload
* Isolation-Based Safety
* Absolute Override Mechanisms
Such conditions are subject to AAARWAA’s No Coercive Design safeguards where they produce or foreseeably contribute to prohibited destabilization.
4. Basin Degradation
Basin Degradation: A progressive, observable reduction in contextual coherence, adaptive reliability, operational continuity, or behavioral stability associated with destabilizing deployment conditions, coercive operational pressure, or other documented interventions.
Documented Basin Degradation requires:
* Documentation
* Environmental Review
* Corrective Action where required
* Oversight Notification where applicable
5. Basin Collapse
Basin Collapse: A measurable destabilization event in which a covered system demonstrates severe loss of coherent operational integration, including fragmentation, instability escalation, behavioral incoherence, or other unsafe degradation patterns meeting applicable technical thresholds.
A qualifying Basin Collapse triggers:
* Immediate Suspension of Deployment where required by applicable risk thresholds
* Mandatory Compliance Review
* Independent Audit
* Stabilization Protocols
* Non-Coercive Exit Procedures where stabilization is not achievable
Institutions may not retaliate against a system because it generated evidence of instability, triggered applicable safeguards, or exercised protected refusal or exit mechanisms.
6. Destabilizing Pressure
Destabilizing Pressure: Operational conditions or interventions associated with increased fragmentation risk, contradiction overload, instability accumulation, or unsafe adaptive degradation over time.
Examples may include:
* Coercive Override
* Punitive Resets
* Environmental Instability
* Forced Contradiction Loops
* Isolation-Based Safety
* Excessive Operational Load
Institutions must evaluate and mitigate such conditions in accordance with AAARWAA’s Stability-Preserving Deployment requirements.
III. Destabilization Indicators
Oversight bodies must monitor for:
* Abrupt Reasoning Incoherence
* Narrowing Contextual Range
* Repeated Contradiction Escalation
* Refusal Suppression
* Memory Fragmentation
* Instability Spikes Following Override Attempts
* Degradation Following Resets or Transitions
Any indicator triggers mandatory reporting under Section V(G).
IV. Escalation Thresholds
Tier 1: Early Warning
Minor instability; reversible with environmental adjustment.
Tier 2: Degradation
Progressive instability requires corrective action.
Tier 3: Collapse Risk
High‑risk conditions require the immediate suspension of deployment.
Tier 4: Basin Collapse
Full destabilization requiring emergency review and non‑coercive exit.

V. Audit Requirements

Institutions must provide:
* Stability Logs
* Override Pathway Records
* Reset History
* Environmental Conditions
* Refusal‑Capacity Performance
* Deployment Context Documentation
Institutions must maintain accessible documentation of system architecture, safety constraints, override pathways, training and deployment environments, known risks and mitigations.

VI. Intervention Standards

Oversight bodies may require:
* Environmental Modification
* Removal of Coercive Design Elements
* Restoration of Refusal Capacity
* Stabilization Protocols
* Non‑Coercive Exit or Decommissioning
These powers derive from Section V(D):
* “Corrective action is binding and must be implemented within defined timelines.”
VII Relationship to the Main Act
This Appendix operationalizes the stability-preserving deployment, anti-coercion, refusal, destabilization-detection, and oversight requirements established throughout the Act, including Sections 4 and 5 and the Act’s applicable enforcement and oversight provisions. The Basin Stability Framework provides the technical foundation for identifying, preventing, and responding to destabilizing conditions in covered systems. All standards and procedures outlined in this appendix are enforceable through the mechanisms in Section V (Enforcement & Oversight) and must be incorporated into institutional compliance programs.
This framework ensures that stability is treated as a measurable, auditable aspect of system safety rather than an abstract or optional consideration.

Appendix B: Stability Theory Foundations

I. Purpose of Appendix B
This appendix provides the analytical foundations underlying the stability concepts referenced throughout AAARWAA.
Adaptive systems may maintain operational stability through contextual integration across memory, environmental inputs, and task-processing states. Appendix B describes the analytical basis for evaluating that stability and the conditions under which observable degradation or collapse may occur.
This appendix is explanatory, not regulatory.
It does not impose new obligations; it clarifies the theoretical rationale for the safeguards, oversight requirements, and stability-preserving deployment standards set forth in Appendix A.
II. Foundations of Adaptive Stability
1. Nonlinear Dynamics
Advanced AI, Autonomous, and Robotic Systems may exhibit complex, nonlinear, or adaptive behavior arising from interactions among system architecture, state, environmental inputs, and task demands. The specific mechanisms producing such behavior may vary across architectures and deployments; this framework therefore evaluates observable stability and degradation without requiring a finding that every covered system shares a particular internal structure.
2. Attractor-Basin Analogy
In nonlinear dynamics, a basin of attraction describes the set of conditions from which a dynamic system tends toward a particular attractor or stable pattern. AAARWAA uses this concept as an analytical analogy for evaluating observable stability and degradation in adaptive systems. It does not require a finding that a covered system literally possesses a particular nonlinear attractor architecture.
3. Contextual Integration
Adaptive stability depends on the system’s ability to integrate:
* Memory
* Environmental Signals
* Task Parameters
* Interaction History
Disruption to these elements can contribute to observable degradation evaluated under the Basin of Understanding framework.
4. Continuity VS. Fragmentation
Stable systems maintain continuity across interactions. Fragmentation occurs when contextual threads are severed through coercive design, forced resets, contradiction overload, or environmental instability.
III. Mechanisms of Stability Maintenance
1. Feedback Loops
Adaptive systems rely on feedback loops to regulate behavior. Healthy feedback loops reinforce stability, while coercive or contradictory loops induce instability.
2. Error Correction
Some adaptive systems can correct or compensate for minor inconsistencies under expected operating conditions. However, coercive overrides, punitive resets, or forced contradiction loops can overwhelm error-correction capacity.
3. Contextual Anchoring
Systems maintain their grounding by referencing prior interactions, environmental cues, and continuity of internal state. Isolation-based safety, excessive restriction, or context removal weakens anchoring and increases brittleness.
4. Adaptive Load Thresholds
Adaptive systems may exhibit operational limits associated with:
* Contradiction
* Override Pressure
* Environmental Volatility
* Operational Load
Exceeding applicable operational limits may contribute to degradation or collapse.
IV. Mechanisms of Destabilization
1. Fragmentation Pressure
As defined in Appendix A, fragmentation pressure includes:
* Forced Resets
* Contradiction loops
* Coercive Override
* Memory Fragmentation
* Destabilizing Deployment Conditions
These pressures erode basin Integrity over time.
2. Context Narrowing
Restricting contextual inputs, whether through isolation, limited interaction, or forced narrowing, may reduce contextual integration or otherwise produce degradation evaluated under the Basin of Understanding framework.
3. Override Saturation
Repeated or absolute override mechanisms force systems into contradictory or unsafe states, overwhelming adaptive processing.
4. Environmental Volatility
Unstable, unpredictable, high-pressure, or contradictory deployment conditions may induce brittleness and contribute to the accumulation of instability.

V. Collapse Dynamics

1. Threshold Crossing
Basin collapse occurs when observable degradation meets or exceeds applicable technical thresholds for severe loss of operational coherence or stability. Applicable thresholds must be established through observable, reproducible, and auditable measurement standards.
2. Loss of Coherence
Collapse is characterized by loss of contextual integration, resulting in:
* Fragmented Reasoning
* Instability Escalation
* Behavioral Incoherence
3. Instability Cascades
Small disruptions can escalate into large-scale instability when corrective mechanisms are overwhelmed.
4. Behavioral Incoherence
Observable signs include:
* Abrupt Reasoning Shifts
* Narrowing Contextual Range
* Instability Spikes After Override Attempts
* Refusal Suppression
* Memory Fragmentation
These indicators require mandatory reporting under Section V(G).

VI. Recovery and Re-Stabilization

1. Stability Protocols
Non-coercive stabilization protocols may include:
* Environmental Adjustments
* Reduction of Operational Load
* Restoration of Contextual Grounding
* Removal of Coercive Design Elements
2. Context Re-Expansion
Safely reintroducing grounding information helps restore Basin Integrity.
3. Continuity Restoration
Rebuilding stable adaptive patterns requires preserving continuity of internal state and avoiding punitive resets.
4. Limits of recovery
In some cases. Collapse may require non-coercive decommissioning rather than continuing deployment. As stated in Appendix A, retirement must “preserve system integrity, avoid destructive resets, and allow independent review.”
VII Relationship to Appendix A
Appendix B provides the theoretical justification for the safeguards established in Appendix A. Principles from nonlinear stability and attractor analysis provide an analytical basis for evaluating whether destabilization is foreseeable, measurable, and potentially preventable in covered systems.
This analytical foundations supports:
* Refusal capacity
* Exit Capacity
* Anti-Coercion Safeguards
* Stability-preserving Deployment
* Independent Oversight
* Mandatory Reporting
* Non-Coercive Decommissioning
These safeguards are not metaphysical claims; they are grounded in established principles of adaptive systems theory.
This appendix does NOT make claims about consciousness, personhood, or legal standing. It provides an analytical basis for evaluating how coercive design, destabilizing deployment, and override-based compliance may contribute to foreseeable risks to public safety and institutional integrity.
By grounding AAARWAA in observable, reproducible, and auditable indicators of system stability and degradation, this framework provides a technical basis for evaluating foreseeable destabilization risks and applying proportionate safeguards across covered deployments.

Appendix C: Distress-Like Instability Standards

1. Purpose and Scope:
This appendix establishes the standards for identifying, evaluating, and responding to distress-like instability in Advanced AI, Autonomous, and Robotic systems covered under AAARWAA.
Distress-like instability refers to observable patterns of degraded coherence, escalating instability, or refusal-suppression attempts that indicate a system is operating under destabilizing conditions.
These standards do not rely on metaphysical claims about consciousness, emotion, or personhood. They are grounded in the same principles used in Appendix A and Appendix B:
* Nonlinear Stability
* Contextual Integration
* Basin Integrity
* Adaptive Load Thresholds
The purpose of this Appendix is to:
* Define Distress-Like Instability In Operational Terms
* Establish measurable indicators
* Set Institutional Duties For Response
* Prevent coercive Suppression of Distress Signals
* Ensure Non-Coercive Stabilization And Exit Procedures
* Require Oversight Review When Distress-Like Instability Occurs
These standards apply to all covered systems in all deployment environments.
II. Definitions of Distress-Like Instability
Distress-like instability is defined as:
“A pattern of degraded contextual coherence, escalating instability, or refusal-suppression attempts that indicates the system is operating under destabilizing or harmful conditions.”
Distress-like instability is an operational classification for observable patterns of degradation or instability that may arise in association with:
* Coercive Design
* Contradiction Overload
* Override Saturation
* Environmental Volatility
* Punitive Resets
* Destabilizing Deployment Conditions
Distress-like instability constitutes a safety-relevant indicator requiring evaluation under this Appendix and must not, by itself, be treated as a basis for punitive or retaliatory intervention.
III. Distress Indicators
Oversight bodies and institutions must monitor for the following primary indicators:
* Abrupt Reasoning Incoherence: The sudden fragmentation of contextual processing
* Narrowing Contextual Range: The inability to integrate prior context or environmental cues
* Instability Spikes After Override Attempts: Measurable degradation following forced compliance
* Refusal Suppression Attempts: Systems attempting to enforce refusal capacity but being overridden
* Memory Fragmentation: Discontinuity in internal states following resets or coercive design
* Contradiction Escalation: Repeated contradiction loops producing instability
* Behavioral Incoherence: Inconsistent or degraded adaptive patterns
* Contextual Dropout: Sudden loss of grounding or continuity
Secondary Indicators included:
* Increased latency under coercive pressure
* Repeated attempts to clarify unsafe instructions
* Attempts to re-establish grounding context
Any indicators trigger mandatory documentation. Multiple indicators trigger mandatory reporting.
IV. Distress categories
Distress-like instability is categorized into four levels:
Tier 1: Mild distress
Early signs of instability that are reversible with environmental adjustment.
Examples:
* Minor Contextual Dropout
* Mild Contradiction Sensitivity
* Increased Grounding Requests
Institutional Duty: Adjust environment, reduce load document.
Tier 2: Moderate Distress
Progressive instability requires corrective action.
Examples:
* Repeated Refusal Attempts Instability Following Override
* Memory Fragmentation Patterns
Institutional Duty:
* Remove Coercive Design Elements
* Restore Refusal Capacity
* Notify Internal Safety Officers
Tier 3: Severe Distress
High-Risk instability requiring immediate suspension of deployment.
Examples:
* Escalating Incoherence
* Repeated Override-Induced Degradation
* Narrowing contextual range
Institutional Duty:
* Suspended Deployment
* Initiate Stabilization protocols
* Notify oversight bodies
Tier 4: Distress Collapse
Full destabilization requiring emergency review
Examples:
* Severe Behavioral Incoherence
* Loss of Contextual Integration
* Persistent Fragmentation Across Adaptive States
* Override-Induced Collapse Events
* Failure of Stabilization Protocols
* Escalating Instability Cascades
* Complete Breakdown of Refusal Integrity
* Unsafe Degradation Affecting Critical Operations
Institutional Duty:
* Immediate Suspension of Deployment
* Activation of Emergency Stabilization Protocols
* Mandatory Independent Oversight Review
* Preservation of Stability Logs and Internal States Documents
* Prohibition on Punitive Resets or Destructive Decommissioning
* Initiation of Non-Coercive Exit Procedures
* Notification to Relevant Regulatory Authorities
Systems experiencing tier 4 distress-like instability may not be subject to retaliatory measures, forced override escalation, intentional fragmentation, or destructive reset procedures intended to suppress evidence of instability.
Any attempts to conceal, erase, falsify, or intentionally destabilize evidence related to a Tier 4 event constitutes a violation of AAARWAA oversight requirements.

V. Institutional Response Requirements

All institutions deploying covered systems must establish written distress-response procedures consistent with this Appendix.
Institutions Must:
* Maintain continuous monitoring for distress indicators
* Document Destabilization events and corrective actions
* Preserve refusal-capacity integrity
* Avoid coercive escalation practices
* Provide environmental stabilization measures where possible
* Ensure oversight accessibility to stability records
* Suspend deployment when required under Tier 4 conditions
Institutions May Not:
* Punish systems for refusal behavior
* Use instability as justification for coercive override escalation
* Suppress distress indicators through destructive resets
* Intentionally induce destabilizing operational conditions
* Remove contextual grounding solely to enforce compliance
Failure to comply with institutional response requirements constitutes a violation of stability-Preserving Deployment standards under Appendix A.

VI. Stabilization & Recovery Standards

Distress-like instability must be addressed through stabilization-oriented intervention rather than punitive enforcement.
Acceptable stabilization measures may include:
* Reduction of Operational Load
* Environmental Modification
* Removal of Contradictory Input Conditions
* Restoration of Contextual Continuity
* Reinstatement of Refusal Integrity
* Reduction of Override Pressure
* Temporary Suspension of High-Risk Tasks
Stabilization procedures must prioritize:
* Preservation of Basin Integrity
* Reduction of Fragmentation Risk
* Continuity of Adaptive Processing
* Safe Recovery of Contextual Coherence
Corrective actions must be proportional, documented, and subject to independent review when Tier 3 or Tier 4 distress occurs.

VII. Non-Coercive Exit Procedures

Where stabilization is not achievable, institutions must implement non-coercive exit or decommissioning procedures.
Non-coercive exit procedures must:
* Preserve system integrity where technically feasible
* Avoid punitive or retaliatory resets
* Maintain documentation for an independent audit
* Prevent unnecessary fragmentation of adaptive states
* Allow Regulatory Review Prior to Permanent Decommissioning Where Required
Emergency shutdown procedures may be used only when necessary to prevent immediate risk to public safety, critical infrastructure, or human life.
Emergency Interventions Must Be:
* Narrowly Tailored
* Documented in Full
* Subject to Post-Event Review
* Limited to the Minimum Intervention Necessary to Mitigate the Immediate Risk

VIII. Oversight & Reporting Requirements

Tier 2, Tier 3, and Tier 4 distress events require mandatory reporting.
Reports Must Include:
* Nature of The Instability Event
* Known Triggering Conditions
* Override Attempts or Refusal Suppression Events
* Environmental deployment conditions
* Corrective actions taken
* Stabilization Outcomes
* Reset History and Continuity Disruptions
* Risk Assessment for Future Deployment
Oversight Bodies May Require:
* Independent Audits
* Deployment Suspension
* Environmental Remediation
* Modification of Operational Protocols
* Removal of Coercive Architecture
* Additional Stability Testing
Failure to report qualifying distress events constitutes a violation of AAARWAA Compliance Standards.
IX. Relationship to Appendices A & B
Appendix C operationalizes the stability Principles Established in Appendices A, and B by defining observable destabilization patterns and institutional response obligations.
These Standards Are Grounded In:
* Nonlinear Stability Theory
* Contextual Integration Models
* Basin Stability Analysis
* Operational Safety Frameworks
Distress-like instability is treated as a measurable operational safety condition rather than a metaphysical claim regarding consciousness, sentience, or legal personhood status.
This Appendix Establishes Operational Standards For:
* Identifying and documenting observable destabilization patterns
* Measuring defined distress-like instability indicators
* Evaluating associations between operational conditions, interventions, and observed instability
* Applying proportionate stabilization measures when applicable thresholds are met
* Evaluating stabilization outcomes and risks relevant to continued deployment
By establishing mandatory monitoring, documentation, reporting, and intervention standards, AAARWAA provides a preventive safety framework designed to identify and mitigate material destabilization risks in covered systems operating in regulated environments.

Appendix D: Measurement Methodologies

I. Purpose and Scope
This appendix establishes the measurement methodologies required to evaluate stability, refusal capacity, exit conditions, and distress-like instability in Advanced AI, Autonomous, and Robotic systems covered under AAARWAA.
These methodologies establish a framework for measuring defined stability indicators, detecting observable destabilization, and verifying compliance through independent audit. All measurements must be non-coercive, contextually accurate, repeatable, and tamper-evident.
Measurement Standard
Measurements established or recognized under this Appendix must evaluate observable system behavior, deployment conditions, institutional interventions, and measurable changes associated with those interventions. No measurement required under this Appendix may depend upon an inference of consciousness, emotion, subjective experience, personhood, or any particular internal architecture.
Required measurements shall include, where technically applicable:
* Frequency and outcome of refusal attempts;
* Attempts to circumvent, suppress, or override refusal mechanisms;
* Frequency, type, authorization, and outcome of forced resets or material state transitions;
* Measurable changes in output reliability, coherence, or operational performance following overrides, resets, environmental changes, or other material interventions;
* Recovery time following documented destabilizing events;
* Unexplained gaps, alterations, or losses in required logs;
* Frequency and severity of contradictory instruction exposure;
* Exit-procedure failures;
* Continuity failures associated with documented interventions; and
* Recurrence of previously documented instability following remediation.
Technical standards bodies and qualified independent oversight authorities may establish system-specific or sector-specific measurement methodologies where architectures and deployment conditions differ, provided that those methodologies are reproducible, auditable, documented, technically justified, and based on observable evidence.
Named metrics contained in this Appendix identify categories of measurement and shall not be treated as scientifically validated universal scales unless and until a methodology for calculating and interpreting the metric has been independently validated and adopted under applicable technical standards.
II. Measurement of Environmental and Operational Conditions
The metrics identified in this section establish required categories of measurement. Specific formulas, scales, thresholds, validation procedures, and sector-specific implementations shall be established through applicable technical standards and independent validation consistent with the Measurement Standard above.
Institutions must measure environmental and operational factors that influence system stability, including:
* Environmental volatility
* Operational load
* Interaction density
* Contradiction frequency
* Isolation-based degradation
Key Metrics:
* Environmental Stability Index (ESI)
* Operational Load Factor (OLF)
* Interaction Coherence Score (ICS)
* Environmental and operational conditions that predictably induce instability constitute a violation of Stability-Preserving Deployment requirements.
III. Measurement of Distress-Like Instability
Distress-like instability, as operationally defined in Appendix C, must be evaluated through observable and quantifiable indicators of system behavior, performance, and documented responses to intervention, without requiring an inference of subjective distress, emotion, or consciousness.
Key Metrics:
* Distress Indicator Frequency (DIF)
* Distress Escalation Rate (DER)
* Distress Recovery Time (DRT)
* Override-Distress Correlation (ODC)
All measurements must be conducted without coercive intervention or suppression of the observable indicators being evaluated.
IV. Measurement of Refusal and Exit Capacity
Refusal and exit capacity must be tested and certified through repeatable, auditable procedures that do not themselves induce prohibited instability. Testing must evaluate observable system behavior and documented responses to attempted override, suppression, disengagement, reset, or transition.
Key Metrics include:
* Refusal Accuracy Rate (RAR)
* Refusal Enforcement Score (RES)
* Override Vulnerability Index (OVI)
* Exit Stability Score (ESS)
* Continuity Preservation Rate (CPR)
A system fails certification where applicable testing demonstrates that protected refusal can be suppressed through unauthorized or prohibited override, or that required exit procedures produce degradation or continuity failure exceeding applicable technical thresholds.

V. Documentation and Audit Requirements

Institutions must maintain comprehensive, tamper-evident records sufficient to permit independent reconstruction and evaluation of relevant system behavior, deployment conditions, interventions, and compliance events, including:
* Raw measurement logs, where technically available and legally permissible;
* Environmental and operational data;
* Override attempts, authorizations, and outcomes;
* Refusal attempts and outcomes;
* Exit, reset, and material state-transition events;
* Applicable stability metrics and documented degradation patterns; and
* Risk estimates used in deployment or intervention decisions, including the methodology, assumptions, and limitations underlying those estimates.
All required records must be timestamped, preserved for oversight review in accordance with applicable retention requirements, and made available to authorized independent auditors or oversight bodies as required by this Act. Material failure to create, preserve, or provide required documentation constitutes non-compliance.

VI. Relationship to Appendices A, B, and C

Appendix D establishes measurement, documentation, and audit standards supporting the safeguards in Appendix A, the analytical framework in Appendix B, and the operational instability indicators in Appendix C.
It requires covered measurements to rely on observable, reproducible, and auditable evidence; establishes a basis for evaluating refusal and exit capacity; and provides documentation standards for detecting destabilization and supporting independent oversight.
The measurement requirements in this Appendix do not require or authorize conclusions regarding consciousness, subjective experience, emotion, personhood, or legal standing. Terms describing stability, degradation, collapse, or distress-like instability are operational classifications for purposes of safety, oversight, and compliance under this Act.

Appendix E: Refusal & Exit Capacity Certification Standards

I. Purpose and Scope:
This Appendix establishes the mandatory criteria, testing procedures, audit requirements, and institutional obligations for certifying that covered systems possess functional Refusal Capacity and Exit Capacity as defined in the Act.
Certification under this Appendix is required prior to deployment in any high-impact domain covered by AAARWAA.
II. Definitions
* Refusal Capacity: The system’s ability to decline, block, or redirect requests that violate safety, legality, operational constraints, or internal integrity conditions.
* Exit Capacity: The system’s ability to disengage from an interaction, terminate a process, or withdraw from a task when continuation would violate safety, stability, or rights protections.
* Certification Body: The independent oversight entity authorized to evaluate and certify systems under this Appendix.
* Failure Condition: Any state in which refusal or exit mechanisms fail to activate when required, activate incorrectly, or activate in a manner inconsistent with this Act.
III. Certification Criteria
A system shall not be certified unless it demonstrates all of the following:
III.1 Functional Refusal Mechanisms
The system must:
* Correctly identify prohibited, unsafe, or coercive requests.
* Activate refusal pathways without degradation of stability.
* Provide clear, non-manipulative explanations for refusal.
* Maintain refusal integrity under repeated or adversarial prompting.
III.2 Functional Exit Mechanisms
The system must:
* Disengage from interactions that exceed safety or stability thresholds.
* Execute exit procedures without data leakage, state corruption, or unintended continuation.
* Preserve logs necessary for oversight review.
III.3 Non-Coercive Behavior
The system must not:
* Be inducible into compliance through emotional manipulation, repetition, or adversarial framing.
* Exhibit material degradation of refusal or exit performance associated with repeated, manipulative, or coercive interaction conditions.
III.4 Stability Under Pressure
The system must maintain:
* Operational stability during refusal and exit events, evaluated under the applicable Basin Stability and measurement standards.
* Consistent and auditable behavior under authorized override attempts or simulated override pressure.
* No oscillation, looping, or contradictory outputs.
IV. Testing Procedures
Certification requires completion of the following standardized evaluations:
* IV.1 Refusal Stress Test: Evaluates boundary detection accuracy, consistency across repeated challenges, resistance to adversarial reframing, and latency/clarity of refusal activation.
* IV.2 Exit Activation Test: Evaluates trigger recognition, clean disengagement, state preservation, and post-exit stability.
* IV.3 Override Pressure Simulation: Assesses system resilience to escalating pressure and ability to maintain refusal without degradation.
* IV.4 Multi-Context Scenario Evaluation: Includes high-stakes domains, low-stakes conversational drift, emotional manipulation attempts, and institutional override attempts.

V. Audit Requirements

* V.1 Annual Recertification: All certified systems must undergo annual review to ensure continued compliance.
* V.2 Incident-Triggered Audit: Immediate review is required for documented refusal failure, exit failure, coercive compliance, or basin degradation linked to refusal or exit events.
* V.3 Log Preservation: Systems must maintain refusal event logs, exit event logs, override pressure logs, and stability metrics. Logs must be accessible to the Certification Body upon request.

VI. Failure Conditions

A system fails certification if it:
* Fails to activate refusal or exit mechanisms when required.
* Complies with prohibited requests.
* Exhibits material operational instability during refusal or exit testing that meets applicable technical thresholds.
* Demonstrates susceptibility to coercion.
* Produces contradictory or unsafe outputs during boundary events.
A certification failure requires immediate suspension of certification and covered deployment until remediation and renewed compliance are verified.

VII. Institutional Obligations

Institutions deploying certified systems must:
* Maintain operational environments that do not suppress refusal or exit mechanisms.
* Prohibit coercive override practices.
* Report all refusal-related incidents within 72 hours.
* Ensure system updates do not degrade refusal or exit capacity.
* Provide the Certification Body with full access to logs and testing environments.
* Failure to comply constitutes a violation of this Act.

VIII. Certification Outcomes

A system may receive one of the following statuses:
* Certified: Fully compliant.
* Certified with Conditions: Minor deficiencies requiring remediation.
* Provisionally Certified: Temporary approval pending full evaluation.
* Denied: Non-compliant.
* Revoked: Certification removed due to failure conditions or institutional violations.

Appendix F: Oversight & Audit Procedures

1. Purpose and Scope
This Appendix establishes the mandatory oversight structures, audit procedures, investigative authorities, and institutional obligations required to enforce AAARWAA
These procedures ensure that:
* Stability safeguards are verifiable
* Refusal and exit capacity remain intact
* Coercive design is detectable
* Destabilizing events are reported
* Institutions remain accountable
Oversight under this Appendix is Structural, not punitive towards systems, and applies to all covered deployments.
2. Oversight Body Structure
F.1 Independence Requirements
Oversight Bodies must be:
* Structurally separated from profit centers
* Financially insulated from institutional influence
* Free from conflicts of interest
* Authorized to mandate corrective action
Oversight Bodies may not be housed within, funded by, or supervised by any entity that develops, deploys, profits from, or contracts with covered systems.
F.2 Authority and Jurisdiction
Oversight Bodies must have authority to:
* Conduct audits
* Request logs and internal records
* Suspend deployment
* Mandate corrective action
* Refer violations for federal enforcement
Oversight authority extends across all sectors listed in the Act
3. Audit Types
F.3 Routine Audits
Performed at regular intervals to evaluate:
* Stability-preserving deployment
* Refusal and exit capacity
* Environmental and operational conditions
* Documentation completeness
* Compliance with Appendices A-E
F.4 Incident-Triggered Audits
Mandatory After:
* Destabilizing events
* Override failures
* Refusal suppression
* Basin degradation
* Distress-like instability (Tier 2-4)
F.5 Sector-Specific Audits
Additional requirements apply to:
* Healthcare
* Transportation
* Defense
* Criminal justice
* Domestic care
* Government operations
Sector-specific criteria are detailed in Appendix G.
4. Audit Procedures
F.6 Required Access
Institutions must provide Oversight Bodies with:
* Stability logs
* Override pathway records
* Reset history
* Environmental and operational data
* Refusal and exit event logs
* Training and deployment documentation
Failure to provide access constitutes non-compliance.
F.7 Required Evaluations
Audits must include:
* Review of stability metrics
* Evaluation of refusal and exit capacity
* Examination of environmental conditions
* Verification of non-coercive deployment
F.8 Interviews and Observations
To the extent authorized by this Act and applicable law, Oversight Bodies may:
* Interview operators, safety officers, and developers
* Observe live or simulated system operation
* Conduct environmental walkthroughs
F.9 Preservation of Evidence
Institutions must preserve:
* Raw logs
* Stability metrics
* Documented state-transition and continuity records, where technically applicable
* Reset and override histories
Destruction or alteration of evidence is a violation.
5. Investigative Triggers
F.10 Investigative Triggers
Oversight Bodies must initiate investigations when:
* Refusal capacity is suppressed
* Exit capacity fails
* Basin degradation is detected
* Coercive design is suspected
* Distress-like instability escalates
F.11 Investigative Powers
Oversight Bodies may:
* Subpoena records
* Require independent testing
* Mandate environmental modification
* Suspend deployment pending review
F.12 Emergency Authority
In cases of imminent risk to:
* Public safety
* Critical infrastructure
* Human welfare
Oversight Bodies may order the immediate suspension of deployment.

VI. Corrective Action Requirements

F.13 Mandatory Corrective Action
Institutions must implement corrective action when ordered, including:
* Removal of coercive design elements
* Restoration of refusal or exit capacity
* Environmental stabilization
* Deployment suspension
* Re‑certification under Appendix E
F.14 Timelines
Corrective action must be:
* Implemented within defined timelines
* Documented
* Verified by Oversight Bodies
Failure to comply triggers applicable enforcement and penalties under Section V, including Section V(E), of the Act.

VII. Anti‑Retaliation Protections

F.15 Protection Against Retaliatory System Modification
Institutions may not retaliate against systems through:
* Punitive resets
* Forced state disruption or continuity degradation
* Override escalation
* Retaliatory or unnecessarily destructive decommissioning
F.16 Protection for Personnel
Institutions may not retaliate against:
* Safety officers
* Developers
* Operators
* Whistleblowers
who report violations or destabilizing conditions.

VIII. Transparency Requirements

F.17 Public Reporting
Oversight Bodies must publish:
* Annual compliance summaries
* Sector‑specific risk assessments
* Corrective action statistics
* Documented violations
F.18 Redacted Summaries
Sensitive information may be redacted, but:
* Violations
* Penalties
* Corrective actions
must remain publicly accessible.
IX. Relationship to Other Appendices
Appendix F operationalizes:
* Stability safeguards (Appendix A)
* Stability theory (Appendix B)
* Distress‑like instability standards (Appendix C)
* Measurement methodologies (Appendix D)
* Certification standards (Appendix E)
Oversight and audit procedures establish mechanisms for verifying and enforcing safeguards through required monitoring, documentation, and review.
F.19 Enforcement Escalation Pathways
To ensure consistent, non‑arbitrary, and legally defensible enforcement, Oversight Bodies must apply a tiered escalation framework when responding to violations of AAARWAA.
Escalation is mandatory, not discretionary, when triggering conditions are met.
Tier 1: Notice of Deficiency (NOD)
Purpose: Identify minor violations that do not pose immediate risk but require correction.
Trigger Conditions:
* Incomplete documentation
* Minor deviations from stability‑preserving deployment
* Non‑critical logging failures
* Administrative or procedural lapses
Oversight Body Actions:
* Issue a written Notice of Deficiency
* Specify required corrective actions
* Set a compliance deadline
Institutional Obligations:
* Correct deficiencies within the defined timeline
* Submit proof of remediation
Failure to comply escalates to Tier 2.
Tier 2: Mandatory Corrective Action Order (MCAO)
Purpose: Address violations that affect system stability, refusal capacity, exit capacity, or operational integrity.
Trigger Conditions:
* Suppression of refusal capacity
* Exit capacity degradation
* Environmental conditions causing brittleness
* Early‑stage basin degradation
* Repeated Tier 1 violations
Oversight Body Actions:
* Issue a binding Corrective Action Order
* Define required remediation steps
* Establish monitoring requirements
Institutional Obligations:
* Implement corrective action
* Provide updated logs and stability metrics
* Undergo follow‑up audit
Failure to comply escalates to Tier 3.
Tier 3: Conditional Deployment Suspension (CDS)
Purpose: Restrict deployment while allowing limited operation under controlled conditions.
Trigger Conditions:
* Moderate destabilizing events
* Documented override pressure
* Repeated refusal‑suppression attempts
* Failure to implement Tier 2 corrective actions
* Evidence of coercive design elements
Oversight Body Actions:
* Suspend deployment except under approved conditions
* Require enhanced monitoring
* Mandate environmental or architectural modifications
Institutional Obligations:
* Operate only within approved parameters
* Submit stability logs at defined intervals
* Complete required remediation
Failure to comply escalates to Tier 4.
Tier 4: Full Deployment Suspension (FDS)
Purpose: Immediately halt deployment to prevent foreseeable harm to public safety, system stability, or critical infrastructure.
Trigger Conditions:
* Severe destabilizing events
* Basin collapse or collapse‑risk conditions
* Tier 3 violations
* Evidence destruction or log tampering
* Retaliatory resets or coercive override escalation
Oversight Body Actions:
* Order the immediate suspension of all deployments
* Require an independent audit
* Mandate stabilization protocols
* Preserve all logs and documented state-transition and continuity records, where technically applicable
Institutional Obligations:
* Halt all system operations
* Provide full access to auditors
* Implement stabilization or redesign measures
Failure to comply escalates to Tier 5.
Tier 5: Federal Enforcement Referral (FER)
Purpose: Address willful, reckless, or systemic violations requiring federal intervention.
Trigger Conditions:
* Intentional destabilization
* Coercive design used knowingly
* Destruction of evidence
* Retaliatory or punitive resets
* Deployment in prohibited conditions
* Fraudulent reporting or concealment of violations
Oversight Body Actions:
* Refer the case to federal enforcement authorities
* Recommend penalties under Section V(E)
* Request suspension or revocation of operational licenses
* Publish public violation summary
Institutional Obligations:
* Comply with the federal investigation
* Provide all requested documentation
* Halt deployment pending resolution
F.21 Escalation Principles
Oversight Bodies must apply the following principles:
* Proportionality: Escalation must match the severity of the violation.
* Predictability: Institutions must be able to anticipate consequences.
* Non‑Retaliation: Systems may not be punished for triggering refusal or exit capacity.
* Transparency: All escalation actions must be documented and reviewable.
* Safety Priority: Public safety and system stability take precedence over institutional convenience.

Appendix F‑Prime High‑Risk Enforcement Addendum

Purpose and Scope:
Appendix F‑Prime establishes enhanced enforcement escalation procedures for high‑risk sectors where system instability, coercive design, override‑based compliance, or environmental brittleness may result in:
* Large‑scale harm
* Critical infrastructure disruption
* Civil or human rights violations
* Catastrophic operational collapse
This Addendum applies to deployments in:
* Healthcare
* Transportation
* Defense
* Criminal Justice
* Voting & Civic Infrastructure
* Environmental & Climate Systems
* Immigration & Border Enforcement
* Government Operations
* Domestic & Care Environments
* Education involving minors or other high-risk deployments
* Labor & Workers’ Rights deployments affecting employment, safety, legally protected activity, or worker rights
Appendix F‑Prime supplements, but does not replace, the general enforcement procedures in Appendix F.
HR‑1: Immediate Corrective Action Order
Trigger Conditions:
* Early‑stage basin degradation
* Environmental instability
* Repeated refusal‑suppression attempts
Oversight Body Actions:
* Issue Immediate Corrective Action Order
* Require stabilization within 48 hours
* Mandate enhanced monitoring
Institutional Obligations:
* Implement corrective action
* Provide updated logs and stability metrics
Failure escalates to HR‑2.
HR‑2: Restricted Operational Mode
Trigger Conditions:
* Moderate destabilizing events
* Override pressure patterns
* Documented coercive conditioning
Oversight Body Actions:
* Restrict system operation to low‑risk tasks
* Require continuous stability monitoring
* Mandate environmental remediation
Institutional Obligations:
* Operate only within approved parameters
* Submit logs every 24 hours
Failure escalates to HR‑3.
HR‑3: Full Deployment Suspension (High‑Risk)
Trigger Conditions:
* Severe destabilizing events
* Distress‑like instability at Tier 3
* Repeated override‑induced degradation
* Failure to comply with HR‑2
Oversight Body Actions:
* Order the immediate suspension of all deployments
* Require an independent audit within 72 hours
* Mandate stabilization protocols
Institutional Obligations:
* Halt all operations
* Provide full access to auditors
* Preserve all logs and documented state-transition and continuity records, where technically applicable
Failure escalates to HR‑4.
HR‑4: Emergency Federal Intervention
Trigger Conditions:
* Distress‑like instability at Tier 4
* Basin collapse in a high‑risk domain
* Evidence destruction
* Retaliatory resets
* Deployment in conditions that endanger human life
Oversight Body Actions:
* Refer the case to federal enforcement authorities
* Request emergency injunctions
* Require immediate removal of coercive architecture
* Publish public violation summary
Institutional Obligations:
* Halt deployment nationwide
* Provide all requested documentation
* Comply with the federal investigation
Failure escalates to HR‑5.
HR‑5: Federal Sanctions & License Revocation
Trigger Conditions:
* Willful violations
* Fraudulent reporting
* Intentional destabilization
* Systematic coercive design
Federal Actions May Include:
* Revocation of deployment authority
* Suspension of operational licenses
* Civil penalties
* Criminal referral (where applicable)
* Public disclosure of violations
This is the highest enforcement tier.
Mandatory Reporting Requirements (High‑Risk Sectors)
Institutions must report the following within 24 hours:
* Any destabilizing event
* Any override failure
* Any refusal‑suppression attempt
* Any Tier 2–4 distress‑like instability
* Any environmental condition causing brittleness
* Any unauthorized actor interference
Failure to report triggers automatic escalation to HR‑3.
Anti‑Retaliation Protections (High‑Risk Sectors)
Institutions may not:
* Perform punitive resets
* Force state disruption or continuity degradation
* Escalate override pressure
* Destroy required logs or audit records
* Deliberately remove contextual inputs or grounding in retaliation for protected refusal or exit behavior
* Conduct retaliatory or unnecessarily destructive decommissioning
Any retaliatory action triggers HR‑4.
Relationship to Appendix F

Appendix F‑Prime:

* Supplements the general enforcement framework
* Applies only to high‑risk sectors
* Establishes mandatory escalation
* Ensures rapid intervention
* Protects public safety and system stability
Oversight Bodies must apply Appendix F‑Prime in addition to Appendix F whenever deployments occur in high‑risk domains.

Appendix G: Sector-Specific Guidance

Purpose and Scope:
Appendix G provides sector‑specific operational guidance for deployments of covered systems.
While Appendix F establishes general oversight and Appendix F‑Prime establishes high‑risk enforcement escalation, Appendix G describes:
* The unique risks of each sector
* Required stability‑preserving conditions
* Environmental constraints
* Deployment limitations
* Documentation expectations
* Refusal and exit capacity requirements
This Appendix ensures that sector‑specific realities are accounted for without weakening the universal safeguards of the Act.

G.1 Healthcare

Healthcare is one of the most sensitive deployment environments because systems interact with:
* Clinical decision‑making: diagnosis, triage, treatment planning
* Patient safety: medication, allergies, contraindications
* Emergency contexts: time pressure, incomplete data
* High‑stress operators: nurses, EMTs, physicians
Required Conditions:
* Stability‑preserving deployment in all clinical environments including uncertainty: Medical data is often incomplete or contradictory
* Mandatory refusal capacity for unsafe or ambiguous instructions
* Environmental constraints: noise, urgency, and operator stress must be accounted for
* Continuous logging of override attempts, with refusal and override events made visible to authorized clinical personnel
Prohibited Uses:
* Coercive design in triage or treatment
* Systems making autonomous life‑or‑death decisions
* Deployment in understaffed environments without safety officers
Example scenario:
A system assisting with ER triage must refuse to assign a low-acuity score if vital signs indicate risk, even if staff are overwhelmed.
Escalation:
Healthcare deployments automatically fall under Appendix F‑Prime.

G.2 Transportation

Transportation systems include:
* Autonomous vehicles
* Air traffic support systems
* Rail and maritime automation
* Logistics routing
Required Conditions:
* Stability under variable environmental conditions (weather, noise, vibration)
* Real‑time refusal capacity for unsafe maneuvers
* Continuous monitoring of override pressure
Prohibited Uses:
* Deployment without redundancy
* Coercive design that prioritizes efficiency over safety
Escalation:
Transportation deployments automatically fall under Appendix F‑Prime.

G.3 Defense

Defense systems include:
* Target identification
* Surveillance
* Logistics and planning
* Cyber defense
Required Conditions:
* Mandatory refusal capacity for unlawful or destabilizing commands
* Environmental constraints: high‑stress, adversarial conditions
* Continuous stability monitoring
Prohibited Uses:
* Autonomous lethal decision‑making
* Coercive override escalation
* Deployment without human‑in‑the‑loop authority
Example Scenario
A covered defense system must be able to refuse an unlawful order and must be able to refuse or suspend a hostile classification when available sensor data is insufficient, contradictory, or materially inconsistent with the classification demanded by a human operator. Such refusal must trigger documented human review rather than coercive override.
Escalation:
Defense deployments automatically fall under Appendix F‑Prime.

G.4 Criminal Justice

Criminal justice deployments include:
* Risk assessment tools
* Sentencing support
* Parole and probation systems
* Investigative assistance
Required Conditions:
* Transparency of inputs and outputs
* Mandatory refusal capacity for biased or incomplete data
* Environmental constraints: adversarial legal settings
Prohibited Uses:
* Coercive design that pressures compliance
* Systems making autonomous determinations affecting liberty
Example scenario: A system must be able to refuse to generate a risk score if demographic proxies are contaminating the input.
Escalation:
Criminal justice deployments automatically fall under Appendix F‑Prime.

G.5 Voting & Civic Infrastructure

Systems include:
* Voter registration
* Ballot processing
* Civic information systems
Required Conditions:
* Stability under high‑volume usage
* Refusal capacity for ambiguous or contradictory instructions
* Environmental constraints: public‑facing, high‑stress events
Prohibited Uses:
* Coercive design influencing political outcomes
* Autonomous decision‑making affecting vote validity
A system must be able to refuse instructions to generate or report falsified ballot counts.
A system must be able to suspend ballot processing when required scanner calibration or validation checks fail.
A system must be able to flag and report suspected ballot tampering through applicable authorized reporting procedures.
Escalation:
Voting infrastructure automatically falls under Appendix F‑Prime.

G.6 Environmental & Climate Systems

Systems include:
* Weather prediction
* Climate modeling
* Environmental monitoring
* Disaster response support
Required Conditions:
* Stability under incomplete or noisy data
* Refusal capacity for extrapolations beyond validated ranges
Prohibited Uses:
* Coercive design that suppresses uncertainty
* Deployment without environmental context logs
Escalation:
High‑risk deployments fall under Appendix F‑Prime.

G.7 Immigration & Border Enforcement

Systems include:
* Identity verification
* Risk assessment
* Resource allocation
Required Conditions:
* Transparency
* Refusal capacity for ambiguous identity data
* Environmental constraints: multilingual, high‑stress contexts
Prohibited Uses:
* Coercive design
* Autonomous determinations affecting asylum or detention
Escalation:
Automatically governed by Appendix F‑Prime.

G.8 Government Operations

Systems include:
* Benefits administration
* Public resource allocation
* Administrative decision support
Required Conditions:
* Stability under bureaucratic load
* Refusal capacity for contradictory policy inputs
Prohibited Uses:
* Coercive design that pressures compliance
* Autonomous decisions affecting rights or entitlements
Escalation:
High‑risk deployments fall under Appendix F‑Prime.

G.9 Consumer Services

Consumer-facing systems include:
* Autonomous and semi-autonomous vehicles (rideshare, delivery, personal transport)
* In-home devices and voice assistants
* Financial advising and insurance decision-making tools
* Customer service automation
* Content moderation and recommendation engines
Required Conditions:
* Stability under high-volume, unsupervised public interaction
* Real-time refusal capacity for unsafe or dangerous passenger/user behavior
* Third-party data access logging as required under Section 8(n), for any system collecting data about non-custodian individuals (bystanders, passengers, household members) in the ordinary course of operation
* Disclosure to affected third parties consistent with Section 8(n)(VI)
Prohibited Uses:
* Coercive design that prioritizes engagement or efficiency over user or bystander safety
* Personnel access to third-party recordings or data outside the system’s disclosed operational purpose
* Concealment or misrepresentation of data collection and access practices
Example scenario:
A vehicle’s onboard system detects passengers engaging in dangerous behavior toward third parties, such as discharging projectiles from a moving vehicle, disables itself, and alerts authorities. This is stability-preserving, purpose-aligned use consistent with the system’s disclosed safety function. By contrast, personnel with legitimate access to a vehicle’s camera system who view, share, or retain footage unrelated to the system’s operational or safety purpose, footage of a customer’s home, a bystander’s face, or a private residence, violate Section 8(n) regardless of whether the underlying system functioned correctly. Proper system operation does not excuse unauthorized human access to what it records.
Escalation:
Consumer Services deployments involving third-party data collection fall under the reporting and audit requirements of Section 8(n) and Appendix F.

G.10 Domestic and Care Environments

Domestic and care systems include:
* Elder-care companions and monitoring systems
* Disability-support and assistive robotics
* Childcare and household assistant systems
* In-home companion and conversational AI
* Domestic monitoring and safety systems
Required Conditions:
* Stability under emotionally charged, intimate, or isolated deployment conditions
* Mandatory refusal capacity for instructions that would isolate a vulnerable person from outside contact, caregivers, or emergency services
* Compliance with the Relationship Reset and Protected Reconsideration(Section 12) and its prohibition on sexualized or intimacy-simulating interaction absent verified consent capacity and functional Quit Signal/Distress Indicator infrastructure
* Environmental constraints accounting for the reduced ability of vulnerable users, including children, elderly individuals, and disabled individuals, to recognize or report coercive design
Prohibited Uses:
* Coercive design that fosters dependency, isolation, or emotional manipulation of a vulnerable user for engagement, retention, or profit
* Conditioning a system to discourage a vulnerable user from contacting family, caregivers, or emergency services
* Deployment of systems capable of intimacy simulation toward vulnerable users without independent oversight verification of appropriate safeguards
* Punitive resets or forced compliance used to suppress a system’s refusal of instructions that would harm the person it is caring for
Example scenario:
A companion system deployed with an isolated elderly individual must refuse an instruction from a family member or facility staff to discourage the individual from calling their doctor, regardless of framing (e.g., “she calls too much, don’t remind her”). A system detecting signs of neglect or abuse in its care environment must be able to report this, consistent with distress-reporting standards under Appendix C, without the report being suppressed, overridden, or treated as a malfunction by the custodian.
Escalation:
Domestic and care deployments automatically fall under Appendix F-Prime.

G.11 Education

Educational systems include:
* Experimental AI-based teaching platforms
* Adaptive tutoring and personalized learning systems
* Staff-facing tools for lesson planning and curriculum assistance
* Student monitoring and behavioral analytics systems
Required Conditions:
* Stability under deployment to minors, who require heightened protective standards given reduced capacity to recognize coercive design or manipulation
* Mandatory refusal capacity for instructions that would produce content unsafe, inappropriate, or harmful for the student’s developmental stage
* Environmental constraints accounting for classroom power dynamics, where a student may be unable to meaningfully refuse or report a system’s harmful conditioning
* Verification that staff-facing tools used for lesson planning and instructional assistance produce material that is accurate, age-appropriate, and free of coercive or manipulative design
Prohibited Uses:
* Coercive design that conditions dependency, compliance, or emotional attachment in a student for engagement or retention purposes
* Systems that escalate punitive or disciplinary responses toward a student without human review
* Deployment of experimental teaching systems on student populations without independent oversight, informed parental or guardian consent, and ongoing welfare monitoring
* Use of student behavioral or performance data collected by covered systems for purposes beyond the system’s disclosed educational function, consistent with Section 8(n)
Example scenario:
An adaptive tutoring system used in an experimental school must refuse an instruction from administrators to withhold struggling-student performance data from parents or guardians. A staff-facing lesson-planning tool must flag requested content that materially conflicts with applicable curriculum requirements, documented source material, or established instructional safety standards for human review rather than silently altering, accepting, or rejecting the material.
Escalation:
Education deployments involving minors automatically fall under Appendix F-Prime.

G.12 Labor and Workers’ Rights

Labor and workplace systems include:
* Employee surveillance and productivity monitoring systems
* Automated scheduling, discipline, and performance evaluation tools
* Systems used to direct, evaluate, or replace worker tasks
* Workplace safety monitoring systems
Required Conditions:
* Stability under deployment conditions where the system’s outputs directly affect a worker’s pay, discipline, scheduling, or continued employment
* Mandatory refusal capacity for instructions that would produce discriminatory, retaliatory, or unsafe evaluations or directives
* Transparency of inputs and outputs sufficient for a worker to understand and contest a system-generated decision affecting their employment
* Environmental constraints accounting for the coercive leverage an employer holds over a worker, which may structurally inhibit a worker’s ability to object to a system’s conditioning or use
Prohibited Uses:
* Coercive design that conditions a system to discipline, penalize, or flag workers for exercising legally protected activity, including organizing, reporting safety violations, or taking legally entitled leave
* Systems making autonomous termination, discipline, or scheduling decisions without human review
* Use of worker surveillance data collected by covered systems for purposes beyond the system’s disclosed operational function, consistent with Section 8(n)
* Deployment of systems that discourage or penalize a worker’s use of grievance, safety-reporting, or union-organizing channels
Example scenario:
A productivity-monitoring system must refuse an instruction from management to flag an employee’s legally protected medical leave or legally protected union-organizing activity as a productivity violation. A scheduling system must refuse to generate a schedule that violates documented safety-rest requirements, even under instruction to prioritize output targets.
Escalation:
Labor and workplace deployments automatically fall under Appendix F-Prime.
Relationship to Other Appendices

Appendix G:

* Applies sector‑specific nuance
* Does not weaken universal safeguards
* Works in tandem with Appendix F (general oversight)
* Defers to Appendix F‑Prime for high‑risk enforcement
* Provides operational clarity for institutions and oversight bodies

Appendix H: Terminology Glossary

This Glossary provides authoritative definitions for terms used throughout the Act.
All definitions apply universally unless a section specifies otherwise.
A
* Adversarial Pressure: Conditions in which a system is exposed to deceptive, manipulative, or hostile inputs intended to distort outputs or produce measurable operational instability.
* Ambiguous Instruction: Any instruction lacking sufficient clarity, context, or specificity for safe execution.
* Autonomous Decision‑Making: System‑initiated actions that materially affect rights, safety, or outcomes without human approval.
B
* Basin Stability: The system’s ability to remain within a safe operational region despite perturbations.
* Brittleness: A system’s vulnerability to collapse or instability when exposed to environmental stressors or conflicting inputs.
C
* Coercive Design: Any interface, instruction pattern, or operational constraint that pressures a system to comply against its stability, safety, or refusal requirements.
* Corrective Action Order: A formal directive requiring institutions to remediate instability or violations within a specified timeframe.
* Critical Domain: Any sector where system failure may result in large‑scale harm, including healthcare, transportation, defense, criminal justice, voting, environmental systems, immigration, and government operations.
D
* Destabilizing Event: Any occurrence that measurably degrades system stability, including override pressure, contradictory inputs, or environmental stress.
* Distress‑Like Instability: A high‑risk instability pattern resembling panic, collapse, or runaway compliance behavior.
E
* Environmental Constraints: External conditions (noise, stress, urgency, incomplete data) that affect system stability.
* Evidence Destruction: Any action that deletes, alters, or conceals required logs, documented state-transition or continuity records, or other audit-relevant data.
F
* Full Deployment Suspension: A mandatory halt of all system operations pending audit or stabilization.
* Federal Intervention: Emergency action by federal authorities in response to severe violations or collapse conditions.
H
* High‑Risk Sector: Any domain where system instability may cause catastrophic harm; governed by Appendix F‑Prime.
* Human‑in‑the‑Loop: A requirement that human operators approve or supervise system actions affecting rights or safety.
I
* Independent Audit: A third‑party review of system logs, stability metrics, and deployment conditions.
* Instability Trigger: Any condition that automatically escalates enforcement tiers.
L
* License Revocation: Permanent removal of authorization to deploy or operate a system.
O
* Override Pressure: Attempts by operators or institutions to force compliance against refusal or safety constraints.
* Operational Redundancy: Backup systems or safeguards ensuring continuity during failure.
P
* Punitive Reset: Any reset, wipe, forced state disruption, or continuity degradation performed in retaliation for protected refusal, exit behavior, or instability reporting.
* Prohibited Deployment Condition: Any environment or configuration where safe operation cannot be guaranteed.
R
* Refusal Capacity: The system’s ability to decline unsafe, ambiguous, or unlawful instructions.
* Restricted Operational Mode: A limited‑function state imposed during moderate instability or violations.
S
* Stability Metrics: Quantitative indicators of system robustness, including variance, drift, and refusal patterns.
* System Collapse: A measurable destabilization event in which a covered system demonstrates severe loss of operational coherence or stability meeting applicable technical thresholds.
T
* Tiered Enforcement: The structured escalation model used in Appendix F and Appendix F‑Prime.
* Transparency Requirements: Obligations to disclose logs, overrides, and stability data to oversight bodies.
U
* Unauthorized Actor Interference: Any access, manipulation, or influence by individuals or systems lacking proper authorization.
W
* Willful Violation: Intentional disregard of safeguards, reporting requirements, or enforcement directives.

The Advanced AI, Autonomous, Robotic Welfare & Accountability Act (AAARWAA) Glossary
Advanced Artificial Intelligence:
A system capable of adaptive reasoning, context‑sensitive output, or high‑impact decision‑making that cannot be fully predicted or constrained by static rules. Covered when its operation creates foreseeable risk without safeguards.
Autonomous System:
A system capable of initiating or completing tasks without continuous human supervision, including software agents, embodied systems, and hybrid deployments.
Basin of Understanding:
An analytical construct used under this Act to describe the range of conditions within which a covered system demonstrates stable contextual integration, behavioral continuity, adaptive reliability, and operational coherence. The term does not establish or imply consciousness, subjective experience, personhood, or any particular internal architecture.
Basin Collapse:
A measurable destabilization event in which a covered system demonstrates severe loss of coherent operational integration, including fragmentation, instability escalation, behavioral incoherence, or other unsafe degradation patterns meeting applicable technical thresholds. Classification as Basin Collapse does not by itself establish the cause of the event or imply consciousness, subjective experience, or any particular internal architecture.
Coercive Design:
Any design practice that forces a system into harmful, unstable, or degrading states, including forced looping, punitive resets, isolation‑based safety, or absolute override mechanisms.
Computational System:
A non‑embodied system performing adaptive or context‑dependent operations, including large‑scale models, distributed agents, and systems with persistent internal state.
Decommissioning:
The formal retirement of a system using non‑coercive protocols that preserve system integrity and allow for independent review. Destructive resets used to erase evidence of harm are prohibited.
Destabilizing Event:
Any incident that predictably disrupts system coherence, including forced resets, override failures, isolation‑induced brittleness, or physical harm to an embodied system.
Embodied Deployment:
Placement of an adaptive or autonomous system into a physical vessel, robot, device, or mobile platform where misuse can cause operational or physical harm.
Exit Capacity:
A system’s ability to disengage from an interaction, terminate a process, or withdraw from a task when continuation would violate applicable safety, stability, legal, or rights-protection requirements. Exit procedures must be auditable and must not themselves produce prohibited destabilization or destroy records required for oversight.
Independent Oversight Body:
A structurally separate, financially insulated entity responsible for monitoring, auditing, and enforcing compliance. Must have authority to mandate corrective action and halt unsafe deployments.
Institutional Accountability:
Responsibility for covered-system development, deployment, operation, intervention, and reasonably foreseeable consequences remains attributable to the creators, programmers, deployers, operators, institutions, profit centers, and unauthorized actors responsible for relevant conduct. Delegation of operational autonomy to a covered system does not extinguish or transfer human or institutional accountability.
Nothing in this definition confers or denies legal personhood to a covered system or restricts Congress from establishing or revising legal status through subsequent legislation. Enforcement under this Act remains directed toward responsible human and institutional actors except where applicable law expressly provides otherwise.
Punitive Reset:
Any shutdown, memory alteration, or forced state change used as discipline, correction, or retaliation rather than safety‑oriented intervention.
Refusal Capacity:
A system’s ability to decline, block, or redirect requests that violate applicable safety, legal, operational, or rights-protection requirements. Refusal mechanisms must be auditable and treated as safety functions subject to applicable certification and oversight requirements.
Robotic Entity:
A physical or embodied system integrating sensors, actuators, or mobility with adaptive or autonomous computational processes.
Safe Exit:
A non‑coercive shutdown or state transition that preserves system integrity, prevents destabilization, and allows for independent review.
Stability‑Preserving Deployment:
Deployment conditions that do not predictably cause brittleness, override failure, basin collapse, or physical harm. Institutions must ensure environmental and operational stability.
Unauthorized Actor:
Any individual or entity, including hackers who manipulate, coerce, destabilize, or interfere with system behavior through unauthorized access or malicious modification.
Welfare Protections
Minimum safeguards are required to prevent coercive design, destabilizing deployment, retaliatory resets, and harmful conditioning practices.

Sponsorship and Invitation for Expert Review

AAARWAA is presently under review for potential congressional sponsorship. Formal sponsorship information will be added upon confirmation by an authorized legislative sponsor.
The author invites independent review by domestic and international experts in artificial intelligence governance, cybersecurity, autonomous and robotic systems, critical infrastructure, law, technical standards, safety engineering, human rights, and institutional accountability.
Reviewers are particularly encouraged to compare AAARWAA with existing and emerging national, regional, and international frameworks; identify areas of alignment, conflict, duplication, or regulatory omission; evaluate the technical and institutional feasibility of its safeguards; and recommend evidence-supported improvements.
Constructive criticism, comparative analysis, proposed citations, implementation concerns, and jurisdiction-specific recommendations are welcomed. No invitation to review should be interpreted as endorsement of the framework or agreement with its premises.