California AI and social media regulation: 26 bills passed

California's Democratic-controlled Legislature passed 26 bills in the final week of its session that would tighten AI and social media use across multiple domains. Governor Gavin Newsom now has until the end of the month to sign, veto, or amend each measure. For AI builders, this is the most significant state-level regulatory wave so far, and it could shape product compliance requirements well beyond California.

What the 26 bills actually cover

The package targets several areas that directly affect how AI products are designed, deployed, and governed:

Minors and social media. One bill would ban "addictive features" such as autoplaying videos for users under 16. Another would impose age verification requirements on AI chatbots used by minors and restrict the ads they can see. Two measures would make platforms and chatbots liable for financial damages if they fail to protect children from harm.

Education. A bill targeting the California State University system would require that instructors be human beings, not AI bots, though instructors may still use AI as a tool in their teaching.

Workplace and surveillance. One bill would ban employers from using AI surveillance products that collect "neural data" to recognize workers' emotional states. Another would bar retailers from using "surveillance pricing" based on dynamic data.

Healthcare. Multiple bills tighten AI use in medical contexts. One requires that therapy and psychotherapy be provided only by qualified licensed professionals and explicitly bars chatbots or companion chatbots from offering those services. Another clarifies that healthcare AI and chatbots remain subject to existing medical confidentiality laws.

Why this matters for AI builders

State Sen. Steve Padilla, an author of one of the bills, described the package as a direct response to the lack of federal action: "While Washington is focused on catering to their tech oligarchs, California is filling the regulatory void." He said the bills could "become the national blueprint" for online child safety. For builders, that means California's rules may influence other states and set de facto compliance standards for any product used by minors.

Compliance costs will rise. If signed, age verification for chatbots, liability for platform harms, and bans on certain AI features will require product teams to invest in identity systems, content moderation, and legal reviews. Products that rely on automated therapy or emotional-analysis features may need significant redesign or face exclusion from the California market.

Human-in-the-loop requirements expand. The CSU instructor mandate and the therapy chatbot ban signal a broader preference for human oversight in high-stakes AI applications. Builders of agentic tools in education, healthcare, or HR should expect similar scrutiny.

Neural data and surveillance restrictions will affect any AI tool that claims to infer emotional states from user data. If your product uses eye tracking, voice analysis, or biometric signals for productivity or engagement metrics, these bills may directly constrain your data collection and processing.

Caveats and open questions

Newsom has not publicly stated which bills he will sign. He has historically governed as a tech industry ally, but recent moves, including an executive order on AI workforce effects and signing earlier AI bills on school services and car tracking, show a more populist turn as he eyes a potential 2028 presidential run. Politico reported that OpenAI CEO Sam Altman directly reached out to Newsom over concerns with some of the measures, but the outcome remains uncertain.

Builders should monitor the governor's decisions closely. Even if some bills are vetoed, the legislature's ambition signals the direction of future regulation. Relying on a single news source for details means that specific bill language and enforcement mechanisms may shift; official texts and subsequent guidance will be essential for final compliance planning.