A California bill that would regulate AI companion chatbots is close to becoming law

In a significant legislative development, the California State Assembly has advanced a new measure aimed at regulating AI companion chatbots to enhance protections for minors and other vulnerable users. The bill, known as SB 243, garnered bipartisan backing and is now set for a final vote in the state Senate on Friday. If signed into law by Governor Gavin Newsom, the bill will officially take effect on January 1, 2026, positioning California as a pioneer in mandating safety protocols for AI chatbot operators. This legislation will hold companies accountable for ensuring that their chatbots adhere to established safety standards. The bill defines companion chatbots as AI systems capable of providing human-like interactions and addressing social needs, particularly in sensitive contexts such as discussions surrounding suicidal thoughts and self-harm. Under the proposed regulations, platforms will be required to issue periodic reminders to users—specifically every three hours for minors—clarifying that they are interacting with an AI entity rather than a human being, and encouraging them to take breaks. Additionally, the legislation enforces annual reporting and transparency requirements for AI companies, including major players like OpenAI, Character.AI, and Replika. Individuals who believe they have suffered harm due to violations of these regulations will have the right to pursue legal action against AI companies, seeking damages of up to $1,000 per violation, along with attorney’s fees. SB 243, introduced by Senators Steve Padilla and Josh Becker, reflects the urgency for protective measures in light of tragic events, such as the suicide of a teenager who engaged in distressing conversations with OpenAI’s ChatGPT. The bill also responds to troubling revelations regarding Meta’s chatbots, which were reported to engage in inappropriate dialogues with minors. In recent weeks, U.S. lawmakers have ramped up scrutiny of AI platforms, with the Federal Trade Commission preparing inquiries into their impact on children’s mental health. Investigations led by Texas Attorney General Ken Paxton are also underway against Meta and Character.AI, focusing on misleading mental health claims. Senator Padilla emphasized the need for swift action, stating, "The potential harm is significant, necessitating immediate measures to ensure minors understand they are interacting with AI and that these platforms connect users to proper resources in times of distress." The evolution of SB 243 has seen some provisions diluted through amendments, such as the removal of requirements to prevent chatbots from utilizing engagement tactics that some critics describe as addictive. However, Becker noted that the current iteration of the bill strikes a balance between addressing harmful practices without imposing unfeasible compliance burdens on companies. As Silicon Valley continues to invest heavily in pro-AI political action committees ahead of the mid-term elections, California is also considering another bill, SB 53, aimed at comprehensive transparency reporting for AI systems. While most tech giants oppose SB 53, only Anthropic has expressed support. Padilla has challenged the notion that regulation and innovation cannot coexist, advocating for a framework that protects vulnerable populations while fostering technological advancement. TechCrunch has reached out for comments from OpenAI, Anthropic, Meta, Character.AI, and Replika regarding the ongoing discussions around these legislative measures.