A new package requires stronger chatbot safeguards, bans some addictive feed features for users under 16, and adds child privacy rules. The next test is whether regulators can make the rules work in real products.
New to this? Read it in simple words
- California signed a large package of online child safety laws. Some ban autoplay and history-based feeds for users under 16.
- Under Adam’s Law, companion chatbots must have crisis steps for signs of suicide or self-harm. Parents must get controls, and a warning if a child turns off safety settings.
- AP reports that large social media companies may face penalties when carelessness causes harm. The penalty can be up to $1 million per child.
- Age checks may ask for more personal data, such as a face scan. Whether children become safer will depend on how well the rules are enforced.
- Companion chatbot
- A chatbot designed to feel like a friend or partner.
- Autoplay
- When the next video or post starts on its own, without a click.
- History-based feed
- A list of posts chosen by what a user watched or clicked before.
The package changes both AI and social media
One law, called Adam’s Law, adds safety duties for companion chatbots used by children. These are chatbots designed to feel like a friend or partner. Companies must add crisis steps for signs of suicide or self-harm, give parents controls, and warn parents if a child turns off safety settings.
The law also requires independent child safety audits and a risk check every year. A risk check asks what could go wrong before or while a product is used. It can help a company find weak points, but only if the test covers real use and the reviewer is truly independent.
Other laws ban autoplay and history-based feeds for users under 16. They also limit targeted advertising, protect school data used by AI, and extend rules on child sexual abuse material to some AI-made or changed images.
California’s new laws cover companion chatbots, addictive feeds, child data, parental controls, and crisis response.
Age checks can create a second privacy problem
A platform must know which users are children before it can apply child rules. That sounds simple, but age checks may ask for an identity document, a face scan, or more personal data. A safety system should collect as little information as possible and delete it when the check is complete.
The rules also need clear product boundaries. A child can move between a chatbot, a game, a social network, and a school tool in one hour. Regulators must explain which duty applies to each service and how companies should work together during a real crisis.
AP reports that large social media companies may face civil penalties of up to $1 million per child when negligence causes harm. A high penalty may push companies to act. It may also lead them to block too much, so courts and regulators will need fair evidence standards.
Watch the audits, controls, and real outcomes
Parents need controls they can understand, not a long settings page. Children also need a clear way to ask for help. A warning should use simple words, show what will happen next, and avoid making a young person feel punished for speaking honestly.
Independent audits should publish the test method, serious findings, and limits. They should measure whether the system notices danger, how often it makes mistakes, and how quickly trained people respond. Company promises are not enough.
The law package sets a strong direction: safer defaults, less addictive design, and more responsibility for companies. Its success will depend on privacy-friendly age checks, honest audits, useful support, and public data showing that harm falls over time.
Sources
Every fact in this story comes from the sources below. Open them to check our work.
- 1Primary source · September 10, 2026Governor Newsom signs the strongest child safety chatbot and social media laws in the nation Office of Governor Gavin Newsom
- 2Research · September 10, 2026California governor signs laws aimed at protecting kids from risks of social media, AI chatbots Associated Press
We used the California governor’s bill summary for the signed measures and AP for independent reporting on penalties, company concerns, and wider context. We do not predict legal outcomes. We focus on the controls people can check after the laws take effect.