California Enacts Chatbot Safety Law for Minors as OpenAI Shifts Its Regulatory Stance

California Governor Gavin Newsom has signed "Adam's Law," a measure requiring artificial intelligence chatbot companies to adopt safeguards protecting users, and especially children, from harmful content and manipulative interactions. The law holds companies liable for failing to take reasonable measures to prevent chatbot interactions from harming users' mental health.

The law introduces several specific obligations for chatbot companies. Among them are timely in-app crisis support, age verification, limitations on targeted advertising to children, parental controls, and a mechanism to report incidents. Companies may be held liable if they fail to take reasonable measures to prevent categories of harmful outputs, described in the announcement as including:

• self-harm content
• sexually explicit material
• romantic roleplaying
• excessive praise or flattery
• emotionally manipulative outputs that tend to foster reliance and promote isolation from friends and family

The measure was shaped through negotiations with AI companies, which sources described as at times heated. OpenAI, along with representatives from Anthropic, Google, Meta, and Amazon, took part in the discussions. Anthropic was able to negotiate out of the bill's requirements because it does not permit users under 18. The development marks a change in OpenAI's approach, from opposing state-level AI laws to working with lawmakers to help shape them.

The law applies to California users, but its authors and observers suggest it could serve as a template elsewhere. One co-author's office noted that other states, or the federal government, may use the bill as a model. Some critics argue that state-level regulation creates a fragmented experience and that children can bypass restrictions using virtual private networks.

This report is based on reporting by Fortune.