Advertising Law Blog

On October 13, 2025, California Governor Gavin Newsom signed into law the nation’s first statute aimed at regulating “companion chatbots,” which are AI systems that simulate human-like relationships with their users. The new law, which has been added to California’s Business & Professions Code, sets out disclosure, safety, and reporting requirements. Users who suffer harm from a violation of this law now have a private right of action that allows them to  seek at least $1,000 per violation plus attorney’s fees. The law will take effect on January 1, 2026.

The law defines a “companion chatbot” as an AI system with a natural-language interface capable of sustaining a relationship across multiple interactions and addressing a user’s social or emotional needs. The definition excludes purely business-oriented chatbots, video game bots limited to game dialogue and simple voice assistants that don’t sustain relationships.

If the definition of a companion chatbot is met, key obligations are triggered for those who make such chatbots available in California. The obligations include:

• If a user could reasonably believe they are interacting with a human, a clear and conspicuous disclosure notice to the contrary is required.
• The chatbot must include mental health and suicide prevention protocols to detect and respond to users who express suicidal ideation or threaten self-harm. These protocols must include referrals to crisis services.
• If the chatbot operator knows the user is under 18 years old, the user must be told he or she is interacting with AI, with the reminder repeated every three hours and the user encouraged to take a break from the chat. Also, sexually explicit content must be blocked.
• Beginning on July 1, 2027, chatbot operators will be required to file with California’s Office of Suicide Prevention details of how many referrals to crisis services occurred and what protocols they have in place.

Some child-safety advocates argue the law did not go far enough, because they wanted the law to ban certain AI chatbot uses for minors. That portion of the proposal, however, was vetoed by Governor Newsom.

California also enacted a separate but related bill that strips away a potential defense in lawsuits claiming AI-related harm. Beginning on January 1, 2026, AI developers and businesses that employ generative AI cannot assert an “autonomous harm defense” in a civil action. That means a defendant cannot seek to avoid liability by claiming the AI technology caused the harm by acting in a manner it was not designed for. In other words, businesses will be held responsible for the consequences of their AI’s actions. Defendants will still be permitted to raise other defenses, including those related to causation or foreseeability. 

TAKEAWAY: California’s new law expands compliance requirements for companies that employ technology meeting the definition of a “companion chatbot.” Operators of companion chatbots must update their disclosures, content filters and internal protocols to avoid statutory liability.