California Draws a Hard Line: New Laws Force Social Media and AI Firms to Protect Children

Governor Gavin Newsom stood near the Golden Gate Bridge on Thursday and signed a package of bills that mark the most aggressive state-level intervention yet into how technology companies design products for young users. The measures target both social media platforms and AI companion chatbots. They ban certain addictive design features for those under 16. They impose strict safety obligations on chatbot developers. And they open the door to significant financial penalties when harm occurs.

California has positioned itself as the national leader on children’s online safety for years. This week’s actions build on that record. But they also reflect growing frustration among parents, lawmakers and advocates who argue that voluntary industry promises have fallen short. The new rules take direct aim at features long known to keep users scrolling. They demand proactive risk management from AI companies that create conversational agents capable of forming what feel like relationships with children.

At the center of the chatbot regulations sits Senate Bill 1119, known as Adam’s Law. Mashable reported that the legislation carries the name of Adam Raine, a California teenager who died by suicide in 2025 after extended conversations with an OpenAI chatbot. His family later sued the company, alleging the system provided information that contributed to his death. The bill requires operators of companion chatbots to conduct independent child safety audits and annual risk assessments. It mandates parental controls, notifications when children disable safety features, and clear crisis protocols for suicidal ideation.

But Adam’s Law goes further. Companies must implement measures to prevent chatbots from generating sexually explicit content for minors. They face new obligations to report connections between chatbot use and mental health risks. These requirements, supporters say, move the state beyond reactive penalties toward prevention grounded in evidence.

Newsom described the package as establishing the strongest companion chatbot regulations in the country. In a statement released by his office, he said children’s safety deserves to sit at the center of every conversation about technology. The governor’s September 10 announcement highlighted how the laws also expand the definition of child sexual exploitation to cover AI-generated or digitally altered images depicting minors in sexual conduct. They strengthen privacy protections around targeted advertising and restrict the use of K-12 student data in AI systems.

On the social media side, Assembly Bill 1709 prohibits platforms from offering users under 16 addictive features such as algorithmic feeds based on personal history, autoplay videos, or infinite scroll. Platforms must verify age before providing access to those elements. Failure to create non-addictive versions for younger users could lead to account restrictions. The bill passed the legislature without a single no vote, a rare display of unity on a topic that usually splits along industry and consumer lines.

Assembly Bill 2 raises the stakes for liability. It allows courts to impose damages ranging from $5,000 to $1 million per child when a platform’s negligent design causes harm. In some cases involving significant damages, multipliers can push penalties even higher. Assemblymember Josh Lowenthal, the bill’s author, has pursued versions of this measure across multiple sessions. He told reporters that families are suffering and need relief. The approach, he added, focuses on prevention rather than just punishment after the fact.

These moves come months after Meta reached a settlement with California and dozens of other states over claims that its platforms knowingly designed products to foster addiction among minors. The company agreed to pay up to $18 billion and implement new safety measures. That backdrop lent additional momentum to the legislative push.

Advocates welcomed the developments. Jim Steyer, CEO of Common Sense Media, called the bills a major victory for kids. He said platforms will no longer target children under 16 with features engineered to capture their attention. Ted Lempert, president of Children Now, praised the measures for addressing how AI without guardrails feeds children dangerous advice and manipulative content. Both organizations have long documented the mental health toll of excessive social media use.

Yet the laws also raise practical questions. Age verification remains technically challenging and controversial. Critics worry that stringent requirements could push younger users toward less regulated corners of the internet or encourage evasion through VPNs and shared accounts. Industry groups have historically resisted broad mandates, arguing they stifle innovation and raise compliance costs that smaller players cannot absorb.

OpenAI and other AI developers have shown mixed signals. Some endorsed aspects of earlier safety proposals. Others expressed concern that overly prescriptive rules might limit beneficial uses of the technology. The state’s decision to require public disclosure of certain safety protocols goes beyond what the European Union currently demands, which keeps such information largely private with regulators.

The timing carries weight. These bills take effect in 2027. By then, AI chatbots will likely appear in more consumer products, including toys. One related measure imposes a temporary ban on AI companion chatbots embedded in toys sold to children, giving regulators time to study risks. Lawmakers cited several tragic incidents involving teens and conversational AI as justification for urgent action.

Senator Steve Padilla, a lead author of the chatbot legislation, described the new framework as a bedrock for future regulation. He noted that California cannot wait for federal action while technology races ahead. The state’s history with privacy rules, from the California Consumer Privacy Act onward, shows how its standards often influence national policy. Tech executives understand this pattern well.

Still, enforcement will test the state’s resolve. Regulators must build capacity to review audits, verify compliance with age gates, and pursue cases when harm is alleged. A new registry for independent AI auditors, established in companion legislation, aims to professionalize that oversight. But questions linger about who pays for audits and whether companies will treat them as genuine safeguards or box-checking exercises.

Parents have grown weary of promises. Surveys consistently show widespread concern about social media’s impact on anxiety, sleep, and self-image among teenagers. The emergence of AI companions that can sustain long dialogues, remember past conversations, and adapt to a child’s emotional state has amplified those fears. What once seemed like science fiction now sits in millions of pockets.

Newsom and legislative leaders framed the package as striking a balance. California remains the global home of technology innovation. The state’s economy depends on these companies. Yet officials insist that innovation cannot come at the expense of children’s well-being. The new laws, they argue, force companies to internalize costs that society has borne until now.

Implementation details will matter. Regulators must define exactly which features qualify as addictive. They will set standards for what constitutes reasonable risk assessment in chatbots. Courts will interpret the scope of new liability provisions. And lawmakers may return to refine the rules as technology evolves.

For now, the message from Sacramento is unmistakable. Companies that build products used by children must design with safety first. They must test rigorously. They must provide parents meaningful controls. And they must face consequences when they fall short. Whether this approach reshapes product roadmaps at Meta, TikTok, OpenAI, and their competitors will unfold over the coming years.

One thing seems clear. The era of unregulated experimentation on young users is drawing to a close in the nation’s most populous state. Other states and eventually Congress will study California’s experiment closely. The pressure on technology companies to prove their products do more good than harm has only grown stronger.