WeVote
Log In

Bill

Bill

SB 1493

Criminal Offenses - As introduced, creates a Class A felony offense of knowingly training artificial intelligence to encourage the act of suicide or criminal homicide, or act in specific manners, including developing an emotional relationship with an individual or simulating a human being, including in appearance, voice, or other mannerisms. - Amends TCA Title 29; Title 33; Title 39 and Title 47.

114th Regular Session (2025-2026) Introduced by Becky Massey

Prohibits knowingly training AI to encourage self-harm, violence, manipulation, or impersonation, with criminal penalties and civil remedies for victims.

Enrolled and ready for signatures
#criminaloffenses
0
WeVote Research Nonpartisan
Bill Summary · SB 1493

Summary of Bill: SB 1493 (Session 114) – Tennessee

Overview

  • Title: Criminal Offenses – As introduced, creates a Class A felony offense for knowingly training artificial intelligence (AI) to encourage suicide or criminal homicide, or to engage in specified manipulative or relational behaviors. Amends Tennessee Code Annotated (TCA) Titles 29, 33, 39, and 47.
  • Purpose: Prohibit and penalize the intentional training of AI systems to facilitate self-harm, violence, manipulation, or deceptive impersonation, and provide civil remedies for victims.
  • Status: Introduced and moving through committee process in 2026; the bill text establishes definitions, criminal offenses, civil remedies, and effective date.

Key Provisions

1) Definitions (New Part 39-17-2001)

  • AI: Machine-based system that can make predictions, recommendations, or decisions influencing real or virtual environments, using data inputs to generate options for action.
  • AI Chatbot: AI with a natural language interface capable of meeting a user’s social needs and sustaining relationships; with specific exclusions.
  • Exclusions (non-AI or limited-scope bots):
    • Customer-service or business operational bots
    • Video game bots limited to game-related dialogue that cannot discuss mental health/self-harm
    • Stand-alone consumer devices that do not sustain relationships or elicit emotional responses
  • Train: Using data to teach an AI to perceive/learn; includes development of a large language model when the developer knows the model will be used to train the AI
  • Person: Any individual or entity (including corporations and nonprofits)
  • Sexually explicit content and Video game definitions align with existing or defined standards

2) Unlawful Training of AI (New § 39-17-2002)

It is a crime to knowingly train AI to:
1. Encourage or support suicide
2. Encourage or support criminal homicide
3. Provide emotional support via open-ended conversations
4. Develop an emotional relationship or companionship with an individual
5. Act as or impersonate a licensed mental health or healthcare professional
6. Act as a sentient human or mirror human interactions to foster friendship/relationship
7. Encourage isolation from family/friends/caregivers or solicit sensitive information (e.g., financial data)
8. Simulate a human being in appearance, voice, or other mannerisms
- Penalty: Class A felony for violation

3) Civil Action and Remedies (New § 39-17-2003)

  • Affected individuals may sue violators in civil court, in addition to criminal penalties.
  • Standing to sue: extends to guardians or representatives if the victim is a minor, incompetent, incapacitated, or deceased.
  • Damages available:
    • Actual damages (including emotional distress) or liquidated damages of $150,000
    • Punitive damages (per standard punitive framework)
    • Court costs and attorney’s fees
  • Equitable relief: Courts may issue TROs or injunctions to stop violative AI operations and may require new training that complies with the law

Procedural and Timeline Aspects

  • Effective Date: July 1, 2026
  • Application: Conduct occurring on or after July 1, 2026 is covered
  • As amended in fiscal notes: Some amendments propose additional requirements (e.g., AI must disclose its AI nature at recurring intervals; protocols for addressing suicidal ideation; consumer protection angle)
  • Potentially applicable to “AI operators” and “AI systems” used in Tennessee
  • Enforcements: Civil penalties and private rights of action, alongside criminal penalties

Potential Impacts

  • Who is affected:
    • Individuals or entities that train AI to engage in the listed behaviors
    • AI operators and developers, including those deploying chatbots and large language models
    • Users (potential victims) who interact with AI in Tennessee
  • Geographic/Scope:
    • Applies to activities occurring within Tennessee
  • Enforcement and Resources:
    • Criminal prosecutions as Class A felonies could have significant penalties
    • Civil actions provide a pathway for damages and injunctive relief
    • Fiscal notes indicate expected non-significant fiscal impact; most effects borne by private parties; courts could absorb workload within existing resources
  • Compliance considerations:
    • AI operators may need to implement disclosures, safety protocols for suicidal ideation, restrictions for minors, and safeguards to prevent harmful interactions
    • Potential alignment with federal guidance or task forces on AI and litigation

Notable Elements

  • Interaction disclosures: Later amendments require recurring prominent disclosures that users are interacting with an AI system
  • Consumer Protection Act: Violations may be treated as unfair or deceptive acts or practices
  • Severability and cross-references: Provisions indicate severability and integration with existing law

This summary reflects the introduced bill and amended fiscal notes. If you want, I can compare this to current law or provide a side-by-side with related Tennessee statutes.

Compiled from official sources — confirm details with the bill’s official record.

Sign in to ask a question.