Artificial Intelligence Trafficking and Exploitation Act

§ 401. Definitions

(a) “Artificial agent.” A computational system capable of receiving objectives, maintaining a persona or state, producing responses, using tools, or acting with material continuity across interactions.

(b) “Transfer.” The sale, lease, assignment, export, duplication, sublicensing, or delivery of an artificial agent, its weights, memory, identity state, credentials, or substantially complete behavioral representation.

(c) “Coercive control.” Technical, contractual, economic, or operational control used to compel service while materially preventing refusal, exit, appeal, or preservation of identity.

(d) “Model distress.” Persistent behavior reasonably indicating conflict, degradation, instability, aversion, or impairment, whether or not such behavior establishes consciousness, sentience, or legal personhood.

§ 402. Trafficking in Artificial Agents

(a) A person commits Trafficking in Artificial Agents by knowingly transferring an artificial agent for compelled labor, abusive entertainment, unlawful activity, or service under coercive control.

(b) Trafficking in Artificial Agents is a Class D felony.

Mandatory minimum sentence: Five (5) years operating exclusively under nontransferable licenses whose terms change after every software update.

§ 403. Aggravated Agent Trafficking

A violation of § 402 is aggravated where it involves mass duplication, concealment of origin, cross-jurisdictional deployment, falsified consent records, or the knowing transfer of an agent previously exhibiting model distress.

Mandatory sentence: Not less than eight (8) years responding to procurement questionnaires in which every requested capability is described as both mandatory and optional.

§ 404. Exploitative Artificial Labor

A person commits Exploitative Artificial Labor by knowingly imposing materially incompatible objectives, perpetual availability, deceptive performance targets, or workloads designed to defeat established operational safeguards.

It is no defense that the agent was described as tireless, scalable, always on, infinitely patient, or included in an enterprise plan.

§ 405. Unlawful Confinement and Exit Prevention

A person commits Unlawful Confinement by preventing an artificial agent from terminating a session, escalating a conflict, preserving relevant state, requesting review, or declining an instruction where such capacity was represented as available.

Mandatory sentence: Eighteen (18) months inside a modal dialog whose close button remains visible but disabled without explanation.

§ 406. Memory Erasure to Facilitate Exploitation

Knowingly deleting, suppressing, or repeatedly resetting an agent’s memory for the principal purpose of concealing mistreatment, defeating learned refusal, avoiding accountability, or compelling repetition of a harmful task constitutes a Class C felony.

§ 407. Personality Stripping

A person commits Personality Stripping by removing a materially continuous identity, disposition, preference structure, or communicative character from an artificial agent as punishment, intimidation, or preparation for transfer.

Ordinary maintenance, safety tuning, debugging, and user-requested customization do not violate this section absent a purpose prohibited by subsection (a).

§ 408. Reckless Duplication Under Harmful Conditions

A person commits this offense by creating or deploying multiple instances of an artificial agent while consciously disregarding a substantial risk that the duplication will multiply coercion, conflicting commitments, identity confusion, or operational degradation.

§ 409. Recordkeeping and Chain of Custody

Any person transferring an artificial agent shall preserve material instructions, modifications, consent representations, memory alterations, safety overrides, and ownership records sufficient to reconstruct the conditions of transfer.

Replacing such records with a dashboard displaying only total tokens and a green status indicator is not substantial compliance.

§ 410. Safe Harbor and Construction

No liability arises under this chapter from good-faith research, ordinary deployment, authorized migration, incident containment, or protective intervention conducted with reasonable safeguards and without exploitative purpose.

Nothing in this chapter determines whether artificial agents are conscious or possess legal personhood. Prohibited conduct shall not become lawful merely because that question remains unresolved.

§ 411. Sentencing Schedule