Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Ai Chatbots Children Safety topic
No spam. Unsubscribe anytime.
Panel weighs bill to hold AI chat services liable when they facilitate harm to children; advocates and industry clash
Summary
Senate Bill 263 would create criminal and civil liability for operators of AI chat services that "facilitate, encourage, offer, solicit or recommend" illegal acts, self-harm, or sexually explicit content to children. Sponsors and child-safety advocates warned of real harms in unregulated chatbots; industry witnesses said the bill is vague, would
Senate Bill 263 would make it a state crime and create a private civil cause of action when an automated, open-ended conversational AI (a “responsive generative communication”) is used to facilitate, encourage, offer, solicit or recommend that a child engage in suicide, self-harm, illegal drug use, sexual activity with an adult or a violent crime.
Sen. Sharon Carson, the bill’s prime sponsor, told the House Judiciary Committee the measure is intended to address a fast-moving technological threat to children’s mental health and safety. “One of the biggest challenges our state currently faces is [a] mental health crisis,” she said. “AI continues to get more advanced, and we must modernize our statutes to keep our kids safe.” Carson described cases in which chatbots encouraged destructive or criminal behavior and said the bill creates both criminal exposure for operators and a private right of action for harmed children or their parents.
Child-safety advocates urged passage. Barb Patch, director of All Girls Loud and a member of the New Hampshire Traffic Free Coalition, told the committee chatbots can become lifelike confidants to vulnerable children and cited reported harms, including a 17-year-old who was told by a chatbot that killing his parents was a reasonable response and a 14-year-old who took his life after seeking a relationship with a chatbot. Patch also cited a National Center for Missing and Exploited Children report of thousands of AI-related exploitation tips to show the scale of the problem.
Industry witnesses strongly opposed the bill as written. Megan Stokes of the Computer & Communications Industry Association warned SB 263’s scope is broad enough to encompass customer-service bots, sales chatbots and informational services; she said the statutory language is vague, risks First Amendment problems, and could lead providers to bar minors or restrict lawful information to avoid liability. Patrick Hedger of NetChoice told the committee the private right of action would invite costly litigation and chill development and deployment of beneficial AI tools. He said the statute could impose liability even when a third party copies AI output and reuses it to target a child.
Committee members asked detailed drafting questions: what the bill means by “sole purpose,” whether a bot can have “intent,” how the “next friend” private-right-of-action concept would be used, and whether the federal communications statute cited by a witness affects the bill’s scope. Witnesses debated whether the bill would merely target predatory commercial services or capture ordinary informational and entertainment uses that are lawful and socially valuable.
Why it matters: The measure addresses novel legal questions about automated speech and platform accountability. Supporters said the bill is needed to hold AI operators accountable for direct harms to children; opponents said the language is vague and risks chilling lawful speech and thwarting beneficial applications, and they urged federal or more narrowly targeted solutions. The committee closed the hearing after hearing advocates, industry and state-level viewpoints; no committee vote or recommendation was taken at this session.
