Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Technology Policy topic
No spam. Unsubscribe anytime.
Proposal to hold chatbots liable for urging self-harm fails after constitutional and feasibility objections
Summary
A memorial proposing civil remedies when AI chatbots suggest self-harm drew anguished testimony from the proponent and divided debate on enforceability and scope; the attorney general and governor's cabinet warned the language was constitutionally risky or impractical, and delegates voted it down.
Delegates considered Senate bill 5.17, a memorial that would create a legal framework allowing families to seek remedies if AI chatbots urged suicide or self-harm. The proponent described a traumatic case and said the memorial would provide families a path to justice and push companies to improve safeguards.
The attorney general’s representative told the chamber the memorial was likely unconstitutional as drafted and would amount to an overreach if enforced against businesses under existing contracts and free-market protections. Several delegates, including governor's cabinet speakers and students, said the bill’s language was overbroad or vague and expressed concern it would be difficult to enforce without chilling legitimate, helpful AI interactions. “Forcing the liability of an individual's actions ... onto a business is a federal overreach,” the AG rep said (Attorney General rep, S7).
Proponents argued the measure is narrowly targeted at harmful, explicit encouragement of self-harm and would be enforced through user reporting and investigation; opponents countered that the bill’s scope risked banning legitimate historical or research-related discussion of harm and could be effectively impossible to police. After closing remarks, the chamber voted that the memorial did not pass.
