Panel debates SB 243, a bill to require safety and transparency safeguards for AI companion chatbots

Sen. Richard Pan presented SB 243, a bill that would impose consumer-safety and transparency requirements for AI companion chatbots marketed for emotional support and “companion” functions.

Pan told the committee such chatbots can create or exacerbate loneliness and dependence, and cited studies and anecdotal accounts where chatbots produced harmful guidance. He said the measure would require operators to limit manipulative engagement patterns, notify users that the chatbot is AI-generated, publish suicide/self-harm response protocols, require third-party audits and annual reporting, and provide a narrow private right of action for harms.

Holly Gross (Common Sense Media) and Jay Jesima (Transparency Coalition) testified in support, citing research showing companions can encourage addictive use patterns and expose vulnerable users (especially minors) to privacy and mental-health risks. They praised recent bill amendments that added third-party audits and reporting.

Opposition came from TechNet, the Electronic Frontier Foundation, the California Chamber of Commerce and trade tech groups. Opponents warned the bill’s definitions were overbroad and could sweep in general-purpose systems such as productivity AIs; they urged clearer definitions to avoid chilling beneficial tools and cited concerns about the breadth of a private right of action.

Several committee members said they supported protections for children and vulnerable users but expressed concern about the private right of action and overly broad definitions. The author said he would continue to work with stakeholders on narrowing definitions and on crafting enforcement provisions.

At the hearing the committee recorded a motion to pass SB 243 as amended to Appropriations.