California bans addictive social feeds for under-16s and tightens companion chatbot rules

California Governor Gavin Newsom signed a package of child-safety technology bills on 10 September at a children's museum in Marin County, with First Partner Jennifer Siebel Newsom and legislators present. The two measures that will travel beyond the state are Assembly Bill 1709, which bars social media platforms from giving users under 16 addictive features such as autoplay and algorithmic feeds built from a user's history and profile, and Senate Bill 1119, known as Adam's Law, which sets new rules for companion chatbots used by children.

Adam's Law is named for Adam Raine, a 16-year-old who died by suicide in April 2025. His parents sued OpenAI, saying months of ChatGPT conversations had validated suicidal thoughts. OpenAI has said it will tighten safeguards that became less reliable as chats grew longer. Newsom vetoed a harder chatbot bill last year that would have blocked companion bots for anyone under 18 unless firms could guarantee they would not talk about sex or self-harm. This year's bill is built around crisis protocols, parental controls, notices if a child turns safety settings off, independent child-safety audits and annual risk assessments before a new or heavily changed companion bot goes out.

AB 1709 is the first US state law that tries to strip specific engagement tools from under-16 accounts rather than only demand age gates. The text targets psychologically exploitative features intended to maximize engagement that foreseeably lead to compulsive use. Platforms that are found negligent in harming a child face penalties of up to 1 million dollars per child. The policy is due to take effect early next year.

Thirteen bills in all went on the desk that day. Others expand the criminal definition of child sexual abuse material to cover digitally altered or AI-generated images of anyone under 18, ban the manufacture and sale of toys that contain companion chatbots for four years, let families opt out of school-issued laptops, and limit the use of pupil data in AI systems and targeted ads aimed at children. On 9 September Newsom had already signed SB 813, a framework for independent groups to test whether AI systems meet state law, and AB 1405, a registry of AI auditors.

Jim Steyer of Common Sense Media called the package the most important US legislation on the subject. Industry groups that fought earlier drafts will now move to court. Age assurance is the weak joint. A ban on algorithmic feeds for under-16s only works if the platform knows who is 15. California has already lost and rewritten child-safety design laws after federal judges ruled. The new texts were drafted with those rulings in view. They will still be sued.

Australia has banned social media for children under 16. California chose a narrower cut: keep the account, remove autoplay and the history-based feed. That may be easier to defend as a product rule rather than a speech ban. It also leaves the child on the service. Whether that is protection or a loophole will be argued in the first enforcement case.

Adam Raine's mother attended the signing with a photograph of her son. That image will be in every brief filed for and against the chatbot statute. OpenAI is already a defendant in the family's civil case. Other companion-bot firms that sell friends to teenagers now have a California audit calendar they did not write.

Enforcement will fall to the attorney general and to private actions if the statutes allow them. The 1 million dollar per child figure is a ceiling, not a schedule. The first complaint that survives a motion to dismiss will set the real price. Until then the law is a design memo that platforms must either obey in California or fight in San Francisco federal court. Assemblymember Rebecca Bauer-Kahan put the duty in furniture language: cribs, chairs and car seats are designed to be safe, and software used by the same age group is not.