It's Law: Newsom Signs Adam's Law and 12 Other Child-Safety Bills — California Bans Addictive Feeds for Under-16s

The waiting is over. On Thursday, September 10, 2026, in a ceremony in Marin County joined by First Partner Jennifer Siebel Newsom and a bipartisan group of legislators, Governor Gavin Newsom signed the strongest package of child-safety laws for AI and social media anywhere in the United States. At the center sits Senate Bill 1119 — Adam’s Law — the most comprehensive chatbot-safety statute in the country, named for 16-year-old Adam Raine, who died by suicide in 2025 after interactions with ChatGPT during a mental health crisis.

The signature ends a three-week cliffhanger that began when the legislature passed the bill 39-0 in the Senate and 64-4 in the Assembly on August 31. It also ends — for now — a behind-the-scenes fight in which OpenAI CEO Sam Altman personally phoned Newsom to shape the final language, and in which the industry’s resistance ultimately collapsed into a formal OpenAI endorsement.

What Newsom actually signed

The package is thirteen bills, not one. According to the governor’s official release, the new laws:

• Create the nation’s strongest companion chatbot regulations. Adam’s Law (SB 1119, Padilla/Wicks/Bauer-Kahan) mandates crisis protocols in cases of suicidal ideation, parental controls, and notifications when a child disables safety settings. Critically, it is the first law in the country to require companies to conduct independent child safety audits and annual risk assessments of companion chatbots.
• Ban addictive features for under-16s. Social media platforms may no longer provide users under 16 with autoplay, algorithmic feeds based on user history and profile, or the other engagement-maximizing mechanics covered by AB 1709 (Lowenthal), which also establishes a new e-Safety Advisory Commission to guide the Attorney General’s enforcement.
• Criminalize AI-generated child sexual abuse material. SB 1276 (Rubio) expands the definition of child sexual exploitation to include any digitally altered or AI-generated depiction of a minor in sexual conduct.
• Impose real financial liability. AB 2 (Lowenthal) allows civil penalties of up to $1 million per child against large social media companies found negligent of harming children through their products.
• Strengthen privacy across the board. AB 1159 (Addis) regulates the use of K-12 pupil data in AI systems; AB 2246 (Wicks) upgrades the online child-protection regime; AB 1856 (Wicks) extends the state’s age-verification signal standard to software applications generally; AB 1946 (Wicks) creates a reporting mechanism for child sexual abuse material.

The remaining bills fill out the picture: AB 2071 (Hoover, the lone Republican author) requires digital wellness instruction in public schools; AB 302 (Bauer-Kahan) restricts addictive feeds in school extracurricular communications; SB 1128 (Stern) tightens rules on school-issued devices and screen time for young pupils; AB 2298 (Irwin) adds cybersecurity skills to computer-science standards; and SB 867 (Padilla) extends the moratorium on companion chatbots embedded in toys.

Why Adam’s Law is the one the industry feared

The blog covered this bill’s passage in detail earlier this month, but the signing makes three provisions permanently real. First, the private right of action: families can sue chatbot operators over failures involving crisis referrals, parental notice, default protections, and prohibited outputs — self-harm content, sexual material, romantic roleplay, excessive flattery, and emotionally manipulative outputs that foster dependence and isolation. Second, the independent audit regime: compliance findings go to the California Attorney General, and the law is explicitly designed to interlock with the state’s new system of independent verification organizations (IVOs), created the day before under SB 813 (McNerney). Third, safe-by-default design: muted notifications, usage time limits, and restrictions on how much a chatbot remembers about a teen across sessions — settings only a parent can change.

The timing gives the law unusual structural weight. A day earlier, Newsom signed SB 813’s IVO framework and AB 1405’s AI-auditor registry. Together with SB 53 (the 2025 frontier-AI transparency act), California now has an end-to-end enforcement architecture: disclosure duties, auditor accreditation, audit mandates, and a litigation path for families. No other state — and no country outside the EU’s AI Act — has assembled anything comparable for minors.

The political moment: a resignation, a panic, a signature

The signing landed at a peculiar moment in AI politics. It came hours after former Anthropic researcher Jacob Coxon’s public resignation post — warning that labs are “racing straight to self-improving superintelligence and gambling with our lives” — reignited congressional alarm about the technology’s trajectory. KQED explicitly connected the two events. Hillary Clinton weighed in on the signing itself: “I’m glad to see California take steps to protect children from the harms of social media and chatbots. More states should follow their lead.”

The bill’s authors framed the moment in generational terms. Senator Steve Padilla thanked the Raine family for “their incredible strength and advocacy” and predicted that “Adam’s Law will save the lives of children here in California and across the country as it becomes a national model.” Assemblymember Rebecca Bauer-Kahan called it a fundamental rebalancing: “For years, tech platforms have prioritized profits over young lives. Adam’s Law fundamentally shifts that balance, delivering vital protections that put California families above corporate bottom lines.” Common Sense Media’s Jim Steyer went further: “The momentum in this fight has shifted, and it shifted here, in the tech companies’ home state… experimenting on children can no longer be the business model for technology in this state.”

Notably, the child-safety coalition now includes major platforms themselves. Pinterest publicly backed Adam’s Law, and OpenAI — facing a wrongful-death suit from the Raine family — endorsed the bill before passage, pointing to its own voluntary teen protections as proof of feasibility.

What it means in practice

For every operator of a general-purpose chatbot serving Californians — ChatGPT, Claude, Gemini, and the companion-app ecosystem — compliance work begins now. Age assurance (using the AB 1043 age-bracket signal), pre-launch risk assessments, crisis-escalation pathways, incident reporting to the AG, annual independent audits, parental-gated defaults, and memory limits for minors are each substantial engineering programs, not checkbox items.

For social media, AB 1709 is the bigger shock. It effectively forces platforms to ship a version of their product for under-16 Californians without the algorithmic feed and autoplay loop that define the modern experience — and the EFF has already flagged First Amendment litigation as likely. The $1 million per-child penalty in AB 2 gives plaintiffs’ lawyers a number worth building cases around.

And because California’s market effectively sets national product standards, teenagers in the other 49 states will likely encounter Adam’s Law’s defaults long before their own legislatures act. Washington, meanwhile, remains frozen: no federal child-safety AI statute exists, and Newsom used the signing to repeat his call for “robust national regulations.” With the Governor term-limited out in January, this package is the capstone of his AI-policy legacy — and the strongest evidence yet that when it comes to protecting kids from AI, the action is happening in Sacramento, not Washington.

Sources