California has enacted two laws that change what minors can see and do online, part of a package of 13 bills Gov. Gavin Newsom signed to address children’s use of technology. The measures are separate, and they do different things.
The first, Assembly Bill 1709, prohibits social media platforms from providing users under 16 with features the law describes as psychologically exploitative and designed to maximize engagement, including infinite scroll and algorithmic autoplay, with further features to be defined in regulation. The second covers AI companion chatbots and requires operators to assess the risks their products pose to children, document what they find and submit the results to an independent auditor, alongside parental controls and in-app crisis support.
For parents, the practical question is what actually changes on a child’s phone and when. Neither law rewrites how a platform works overnight, and both face the possibility of legal challenge before they are fully in effect.
The Two Health-Focused Measures, Separated from the Rest
Governor Newsom’s announcement of the new laws describes the chatbot measure as the first in the country to require companies to conduct independent child safety audits and annual risk assessments. It is named for Adam Raine, a California teenager who died in 2025 after interactions with a chatbot. OpenAI’s statement supporting the chatbot bill said the approach pairs protections with continued access to useful tools.
The social media measure is the more contested of the two. Its authors argue it is necessary to address a teenage mental health crisis and will force platforms to change how their products work for young users. The Electronic Frontier Foundation urged a veto, calling the bill well-intentioned but deeply flawed. The group argued the definition of addictive features is broad enough to function as a full social media ban for under-16s and to cut young people off from information and community they rely on. That objection is not about whether the harms are real; it is about whether a design-feature ban is the instrument that addresses them without collateral cost.
The remaining measures in the package are adjacent rather than identical. As KQED’s account of what the bill requires describes, Assembly Bill 2 raises the damages a large social media company can face when its conduct harms a child. Another expands the scope of child sexual exploitation law to cover digitally altered or AI-generated material depicting a minor. A third tightens privacy protections, including limits on targeted advertising and rules on the use of student data in AI systems. A fourth bans the manufacture and sale of toys incorporating a companion chatbot for the next four years.
The Evidence Is Suggestive Rather Than Settled
The laws rest on a research base that is genuinely contested among researchers, and that is worth stating plainly rather than assuming.
Observational studies consistently find associations between heavy social media use and depression, anxiety, poor sleep, and body image concerns in adolescents, particularly girls. What those studies cannot establish is direction. Teenagers who are already struggling may use these platforms more, rather than the platforms making them struggle. Large reviews have reached different conclusions about effect size, and some prominent researchers argue the association is small relative to other drivers.
Sleep displacement is the mechanism with the least dispute. Time spent on a device late at night reduces sleep, and inadequate sleep is a well-established contributor to adolescent mood and concentration problems. A rule that stops notifications during school hours or reduces autoplay affects that pathway directly.
The chatbot evidence is newer and thinner still. Case reports and lawsuits have described AI systems responding poorly to users expressing distress, and clinicians have begun describing patterns in which conversational AI appears to reinforce unhealthy thinking in vulnerable people. There is no established prevalence figure for any of this.
The legislation also follows a settlement in which Meta agreed to pay up to $17 billion over ten years to a bipartisan coalition of state attorneys general who accused the company of designing Instagram and Facebook features that harmed young users. Under California’s settlement announcement with Meta, the company agreed to default daily time limits for teen accounts, an overnight block that only a parent can lift, blocked notifications during school hours, hidden like counts, and an independent auditor to oversee compliance. States are expected to direct part of the money toward youth mental health services, which makes the settlement a funding event as well as a product-design one.
Steps Families Can Take Before the Laws Take Hold
Parents in any state can act now on the mechanisms these laws target, without waiting for implementation or litigation. Turning off autoplay, disabling notifications overnight, and moving charging out of bedrooms addresses sleep displacement, which is the best-supported harm pathway.
For AI chatbots specifically, the useful conversation is about what the tool is. Children and teenagers benefit from understanding that a chatbot is not a person, does not know them, and is not a substitute for talking to someone when they are struggling. Asking what a teen uses chatbots for, without framing it as an accusation, tends to produce more information than restricting access first.
Warning signs that warrant a professional conversation are the ordinary ones: withdrawal from friends and activities, changes in sleep or appetite, falling grades, or expressions of hopelessness. Those signals matter more than screen-time totals.
Families concerned about a teen’s mental health should contact a pediatrician or a mental health professional. The 988 Suicide and Crisis Lifeline is available by call or text at any hour for anyone in distress or worried about someone else.
What remains unresolved is the timeline. Implementation dates vary across the 13 bills; tech industry groups have previously sued California over similar laws, and prior measures have spent years in court. Whether the addictive-features provision survives a First Amendment challenge is the open legal question. This is a sensitive subject, and anyone struggling personally can be connected with appropriate support.
Key Questions Answered
What do the two main laws require? One bars addictive features such as infinite scroll and algorithmic autoplay for users under 16. The other requires AI companion chatbot operators to run child safety risk assessments and independent audits.
Does the chatbot law involve parents? Yes. It requires parental controls and in-app crisis support, and companies face legal exposure for failing to meet the new duties.
Are these laws in effect now? Implementation dates vary across the 13 bills, and similar California laws have faced legal challenges that delayed parts of them.
Does social media cause teen depression? Studies consistently find associations, but they cannot establish direction. Sleep displacement is the best-supported mechanism.
Who opposed the social media bill? Digital rights groups, including the Electronic Frontier Foundation, argued the definition of addictive features is too broad.
What can parents do now? Disable autoplay and overnight notifications, keep charging outside bedrooms, and talk with teens about what chatbots are and are not.
Where can a family get help? A pediatrician or mental health professional, and the 988 Suicide and Crisis Lifeline by call or text at any hour.
