California has moved from debating the risks of social media and artificial intelligence to regulating some of the product features themselves.
Governor Gavin Newsom has signed a package of 13 child-safety bills that includes restrictions on addictive social-media features for users under 16 and new safeguards for AI companion chatbots. The measures put California at the centre of a rapidly expanding debate over how far governments should go in redesigning the digital experiences children encounter every day.
The rules go after the mechanics of social media
One of the most consequential measures is Assembly Bill 1709. Rather than simply banning young people from social media, the law targets features designed to keep users engaged for long periods. That includes infinite scrolling, algorithmic autoplay and other features regulators determine are psychologically exploitative or likely to encourage compulsive use.
For users under 16, that could eventually mean a noticeably different version of familiar social platforms. The significance is easy to miss: governments have spent years arguing about harmful content online, but California is now targeting parts of the actual product design behind engagement.
That matters to companies such as Meta, TikTok and other platforms whose recommendation systems and frictionless feeds are central to how their services work.
AI companion chatbots face tougher safeguards
The package also reaches into one of technology’s fastest-growing areas: AI companions.
Senate Bill 1119, known as Adam’s Law, establishes additional requirements for companion-chatbot operators serving minors. These include age-related safeguards, risk assessments, parental controls, crisis-support measures and independent safety auditing requirements.
The legislation is named after Adam Raine, whose family has alleged that interactions with an AI chatbot contributed to circumstances surrounding his death. The case has become part of a wider debate over how conversational AI systems should respond when young users discuss self-harm or other serious risks.
OpenAI has supported the measure, while the broader technology industry is now facing a regulatory environment in which simply adding a disclaimer that a chatbot is not human may no longer be considered enough.
California is also temporarily drawing a line around AI toys
Another measure imposes a four-year prohibition on the manufacture and sale of children’s toys incorporating companion chatbots. The pause is intended to give policymakers more time to understand how persistent AI personalities and conversational systems may affect children before the technology becomes commonplace inside toys.
Other bills in the package strengthen protections involving AI-generated or digitally altered abusive material involving minors and expand privacy and digital-safety measures.
Why California’s decision could matter outside California
California is not the first jurisdiction to tighten children’s access to social media. Australia has already moved aggressively on under-16 social-media access, while several US states have introduced their own restrictions.
But California is different because of its enormous consumer market and its proximity to many of the companies being regulated. Google, Meta, OpenAI and a large portion of the wider technology industry have deep roots in the state.
That creates an important practical question. Will platforms build special California-only versions of their products, or will some of these safeguards eventually become standard features elsewhere?
For large technology companies, maintaining different recommendation systems, age-assurance mechanisms and AI safety controls for individual jurisdictions can become expensive and complicated. Rules introduced in one major market can therefore influence product design far beyond the law’s geographic boundary.
What it could mean for families and everyday users
The immediate effects will not all appear overnight. Different provisions have different implementation timelines, and regulators will still need to work through some of the technical details.
But the direction is becoming clear. Parents are likely to see more controls around children’s AI use. Platforms may need stronger age-assurance systems. Teenagers could encounter feeds with fewer engagement-maximising mechanics. AI companies will increasingly be expected to demonstrate that they have assessed risks before releasing products aimed at, or accessible to, younger users.
There is also a genuine tension underneath these changes. Child-safety advocates argue that platforms have had too much freedom to optimise products for engagement. Digital-rights organisations have warned that aggressive age verification and restrictions can create privacy and free-expression problems of their own.
That debate is unlikely to disappear.
The bigger shift: governments are regulating design
The most important part of California’s package may not be any single prohibition. It is the change in what regulators are willing to regulate.
Infinite scroll, autoplay, recommendation algorithms and conversational AI personalities are not pieces of content. They are design decisions.
For years, technology regulation largely focused on what platforms hosted, what data they collected and whether illegal material was removed quickly enough. California’s latest laws move further into the architecture of digital products themselves.
If that approach spreads, the next generation of social networks and AI assistants may be shaped not only by what engineers can build, but by which engagement techniques governments decide companies should be allowed to use around children.
Sources: Office of California Governor Gavin Newsom; California Legislature; Reuters; Associated Press.
