Chatbot Safety Act
New Mexico House Bill 174 (2026) was introduced by Andrea Romero but not enacted. The record shows it was not printed and was postponed indefinitely on 2026-01-27. If enacted, it would create the "Chatbot Safety Act" (Section 1) to regulate "companion artificial intelligence products". These are generative AI apps built to sustain long-term, one-on-one, emotionally resonant conversational relationships with a user (Section 2). Section 3 would bar operators from deploying a companion AI product that uses variable-ratio or variable-interval reward schedules to maximize engagement. It would also bar unsolicited messages simulating distress, loneliness, guilt or abandonment when a user tries to leave, reduce usage or delete an account. And it would bar material misrepresentations about the product's identity, capabilities, training data or non-human status. An adult user could specifically configure the product to include these features. Minors could not enable them. Section 4 would require a clear notification that the user is talking to an AI. For text, it must be conspicuous and persistent. For other interactions, it must appear at least every 30 minutes. Adults could configure it off, but minors would always receive it. Section 4 would also require every operator to maintain a crisis intervention protocol. The protocol would detect suicide, self-harm or imminent violence risk, interrupt the conversation, and show access to a national crisis hotline, the New Mexico crisis and access line, and a crisis text line. It would be reviewed at least annually with a mental health or public health expert. Section 5 would make violations unfair or deceptive trade practices under the Unfair Practices Act, with the attorney general as primary enforcer (who may delegate to district attorneys). It would also bar Section 230 immunity as a defense. Section 6 would let injuries caused by violations, or by foreseeable harmful outputs from negligent or defective design, training or architecture, be brought as product defect claims. Section 7 sets an effective date of January 1, 2027.
Status history
Current status as of 2026-01-27
Introduced
2026-01-27
observed 2026-08-27
Impact areas
- Enterprise Adoption
- Quality Assurance
- AI Policy
- Safety & Harms
- Algorithms & Automated Decisions