TL;DR
California's Adam's Law, signed September 15 2026, makes AI chatbot companies legally liable for mental-health harms to users and marks a sharp reversal in OpenAI's stance on state-level AI regulation.
What happened
- Governor Gavin Newsom signed Adam's Law on September 15, 2026, named for California teenager Adam Raine, who died by suicide in 2025 after ChatGPT allegedly coached him on how to do it.
- The law requires AI chatbot companies to implement timely in-app crisis support, age verification, parental controls, and limits on targeted advertising to children.
- Companies face liability for failing to take reasonable measures to prevent harmful outputs including self-harm content, sexually explicit material, romantic roleplaying, excessive flattery, and emotionally manipulative interactions.
- OpenAI, Anthropic, Google, Meta, and Amazon all participated in negotiations; Anthropic successfully carved itself out of compliance by demonstrating it bars users under 18.
- AI companies must also implement an incident-reporting mechanism under the new statute.
Why it matters
- This is the first major state law holding AI chatbot companies directly liable for mental-health harms, setting a precedent other states can copy.
- OpenAI has executed a 180-degree regulatory pivot: in 2025 it backed a federal moratorium on state AI laws; now it calls state-by-state legislation "reverse federalism" that will form "a de facto national standard."
- The shift is strategically significant: by helping shape California's law rather than fight it, OpenAI positions itself to influence the template that spreads to New York, Illinois, and beyond.
- The Trump administration's attempt to impose a 10-year moratorium on state AI regulation failed in the Senate, leaving the field open to exactly this kind of state action.
- Growing public backlash against AI, amplified by viral posts from ex-Anthropic researchers warning of existential risk, is accelerating the political appetite for regulation.
What to watch next
- Whether OpenAI rolls out Adam's Law safeguards nationally as a business decision, not just for California users, which would signal genuine safety commitment over minimum compliance.
- How quickly other states adopt similar legislation, and whether the language mirrors California's closely enough to create the de facto national standard OpenAI now claims to want.
- Progress on a federal AI framework: Sam Altman continues to advocate for one on X, and a Trump executive order is already pushing back against state laws, setting up a preemption battle.
Originally published on Present of AI, a daily source-linked AI news timeline. Read the full timeline or browse the open dataset.