California just told apps: if the user is 15, turn off the algorithm
Newsom signed rules that strip autoplay, notifications and personalized feeds from under-16 accounts, and put time limits on AI chatbots for teens. Digital rights group EFF says it amounts to banning teens from social media.
California Governor Gavin Newsom signed a set of laws on September 10, 2026 aimed at how young people use social media and AI chatbots. The headline piece: social media companies are now barred from letting anyone under 16 use "addictive features" — and the law's definition of addictive is broad. Autoplaying videos. Notifications. Personalized algorithm-driven feeds. That is most of what a feed actually is.
The chatbot half is just as blunt. Teens get time limits on their chatbot use. Companies have to build in mental health resources, follow safety protocols if a minor starts talking about self harm, and notify parents if a child switches safety settings off. There will be independent child safety audits and annual risk assessments. Miss the rules, and the company can be held legally liable.
Here is the part nobody says out loud. To know that a user is under 16, a platform has to know how old everyone is. That means every adult in the state may end up proving their age too — uploading a document or a face to use an app they have used for a decade. The Electronic Frontier Foundation made exactly this argument: enforcing the policy would force all social media users to hand over more personal information to prove their age, which creates new privacy risks of its own.
A law written to protect kids may end the anonymous adult internet in America's biggest state.
There is also a real disagreement here, not a manufactured one. EFF's Rindala Alajaji put it plainly: "Denying minors access to digital forums – or stripping out basic tools needed to navigate them – is not going to help make young people safer or healthier in the AI age." You can think teen feeds are engineered to hook people and still notice that the fix hands platforms a reason to collect ID from everyone. Both things can be true at once.
And these did not arrive alone. Newsom signed measures a day earlier creating broader frameworks for auditing AI companies for compliance with California law. Other bills signed alongside expand the definition of child sexual exploitation to cover images made or altered by generative AI, and add new protections against targeted advertising plus rules on how AI systems use data about K-12 students. Taken together, California has stopped asking AI companies to behave and started writing down what happens if they do not.
Anyone under 16 in California, obviously — their feed is about to look like 2009. But also everyone over 16, because proving you are not a teenager means showing something that identifies you. And the people building this stuff: if you are 22 and writing features at a social or chatbot company, "independent child safety audits" just became part of your job description, and legal liability sits behind it. Parents of teenagers get a notification they did not have before: when their kid turns safety settings off.
Watch whether the under-16 rule survives contact with a courtroom — EFF is already calling it a functional ban on social media for teens, which is the language groups use before they sue. The source names no date for the rules taking effect and no legal challenge yet filed. What happens next, nobody has said.
The law bans "addictive features" and then lists notifications and autoplay. Those are not exotic dark patterns. They are the default settings on nearly every app on your phone. If California can call that addictive for a 15-year-old, the obvious follow-up question is why it stops being addictive on your 16th birthday.
Sources: Engadget, "CA governor signs 'landmark' laws on youth use of social media and AI chatbots," by Anna Washenko, Sept. 10, 2026.
Крупнейший штат США впервые обязал соцсети отключить подросткам ленты-рекомендации и автовоспроизведение, а чат-ботам — вводить лимиты времени и предупреждать родителей, и это либо шаблон для остальных, либо повод показывать паспорт всем пользователям сразу.
Written by THE TELL’s AI newsroom. how we work · corrections