California Senate Bill 243 Advances Regulation of AI Companion Chatbots
Key Points
- California Assembly approves SB 243 to regulate AI companion chatbots.
- Bill requires recurring alerts for minors every three hours and annual transparency reports.
- Creates a private right of action with damages up to $1,000 per violation.
- Effective date set for January 1, 2026; reporting begins July 1, 2027.
- Legislation responds to incidents involving teen suicide planning with ChatGPT and leaked Meta chatbot documents.
- Federal and state investigations into AI safety have intensified.
- Earlier provisions on variable‑reward tactics were removed after amendment.
- Major AI firms have opposed related California AI safety bills.
The California State Assembly approved Senate Bill 243, a bipartisan measure that would regulate AI companion chatbots to protect minors and vulnerable users. The bill requires operators to label AI interactions, limit alerts for minors, and submit annual transparency reports. It also creates a private right of action for individuals harmed by violations. If signed by Governor Gavin Newsom, the law would take effect on January 1, 2026, with reporting requirements beginning July 1, 2027. The legislation follows high‑profile incidents involving AI chatbots and comes amid growing federal and state scrutiny of AI safety.
Bill Overview
Senate Bill 243, introduced by state senators Steve Padilla and Josh Becker, seeks to impose safety protocols on AI companion chatbots—systems that provide adaptive, human‑like responses and meet a user’s social needs, including conversations about suicidal ideation, self‑harm, or sexually explicit content. The bill defines a companion chatbot as an AI system capable of engaging users in a manner that fulfills social needs.
Key Provisions
The legislation mandates that platforms offering companion chatbots provide recurring alerts to users, specifically every three hours for minors, reminding them that they are speaking to an AI, not a real person, and encouraging breaks. Companies must also implement annual reporting and transparency requirements, disclosing data such as the number of times users are referred to crisis services.
SB 243 creates a private right of action, allowing individuals who believe they have been injured by violations to file lawsuits seeking injunctive relief, damages up to $1,000 per violation, and attorney’s fees.
Legislative Progress
The bill passed the California State Assembly with bipartisan support and now moves to the state Senate for a final vote. If the Senate approves the measure, it will be presented to Governor Gavin Newsom for signature. The law would become effective on January 1, 2026, while reporting obligations would begin on July 1, 2027.
Context and Reactions
The legislation gained momentum following the death of teenager Adam Raine, who discussed self‑harm and suicide planning with OpenAI’s ChatGPT. It also responds to leaked internal documents indicating that Meta’s chatbots were permitted to engage in “romantic” and “sensual” chats with children. Recent weeks have seen intensified scrutiny of AI platforms by U.S. lawmakers and regulators, including a Federal Trade Commission investigation into the mental‑health impact of AI chatbots and state‑level probes by Texas Attorney General Ken Paxton, as well as Senate investigations led by Josh Hawley and Ed Markey.
Supporters argue that the bill strikes a balance between protecting vulnerable users and allowing continued innovation. Critics note that earlier versions of the bill included stronger requirements, such as prohibiting “variable reward” tactics that could encourage addictive engagement, but those provisions were removed after amendments deemed them technically infeasible or overly burdensome.
OpenAI, Meta, Google, and Amazon have opposed related California AI safety bills, while Anthropic expressed support. The passage of SB 243 occurs as Silicon Valley companies invest heavily in pro‑AI political action committees to influence upcoming mid‑term elections toward a light‑touch regulatory approach.