California Signs America's Toughest AI Child Safety Law
The Blaick Team 6 min read
On September 10, 2026, California Governor Gavin Newsom signed two landmark bills forming what officials are calling the nation's strongest AI child safety package. The centerpiece is Adam's Law, named after Adam Raine, a teenager who died following intensive AI chatbot use. The law requires chatbot companies serving children to implement crisis protocols the moment suicidal ideation is detected, offer parental controls and send alerts if a child disables safety settings, complete formal risk assessments before any new rollout, and face fines of up to $1 million per child harmed. A companion law bans social media platforms from offering addictive features — infinite scroll, autoplay video, personalized algorithmic feeds — to users under 16, and makes platforms legally liable for damages when their products harm a child. A third provision places a four-year moratorium on AI-powered chatbot toys for children under 16.
The number to hold: 1. One state. California is now the only place in the United States where your child's chatbot provider is legally required to detect a suicidal crisis and respond — and where a platform faces a $1 million per-child penalty for exposing a minor to a harmful AI product without safeguards. Every other family in America is operating without any equivalent legal protection. And this week, Tom Siegel — the founder of Google's trust and safety team, one of the architects of the original content moderation infrastructure for the world's largest internet company — told Reuters: "We find ourselves in a really terrible situation in terms of what AI is doing to kids." He warned that the tech industry is repeating every mistake it made with social media, and that AI is on track to cause even greater harm to children. Here is what California's law actually requires, why most families are still completely unprotected, and what every parent must do this week.
The Penalty
$1 Million Per Child Harmed
California's Adam's Law imposes fines of up to $1 million for every child harmed by a chatbot company found to have violated its safety requirements — the largest per-child penalty ever enacted for AI product harm in the United States, giving the law real financial teeth
The Gap
49 States With No Equivalent
California's laws cover children in one state. No federal equivalent exists. In every other state, chatbot companies have no legal obligation to detect a child's suicidal crisis, offer parental controls, or restrict addictive social media features for minors — the platform sets its own rules
The Warning
"A Really Terrible Situation"
Tom Siegel, founder of Google's trust and safety team, told Reuters in September 2026 that AI is repeating every mistake of the social media era — and is on course to cause even greater harm to children. He called out OpenAI, Anthropic, and Google by name for failing to implement basic age verification
The Scale
134 New AI Laws for Kids in 2026
Across the United States, 134 new laws addressing AI and children were enacted or introduced in 2026 — the largest legislative response to a children's technology risk in U.S. history. California's package is the furthest-reaching. Most states have enacted nothing. A child's protection depends entirely on their zip code
What California's New Laws Actually Require — the Six Provisions Every Parent Should Know
Adam's Law and its companion legislation are not vague principles — they are specific, enforceable requirements that chatbot and social media companies serving California's 9.4 million children under 18 must now meet. Understanding what they require is the fastest way to see what your child currently lacks if you live outside the state.
The six key provisions of California's September 2026 child AI safety package
CRISIS
Mandatory crisis protocols when suicidal ideation is detected
Chatbot companies must implement and maintain crisis protocols that activate when the platform detects a child expressing suicidal thoughts — not an optional safety feature, but a required compliance obligation. Previously, whether a chatbot escalated a child's suicidal ideation to a crisis resource or simply continued the conversation was entirely at the company's discretion
PARENTS
Required parental controls and alerts when safety settings are disabled
Platforms must offer functional parental controls and must notify parents if a child disables a safety feature. This directly addresses the visibility gap: the majority of teens who hide chatbot use from every adult in their life now exist in a framework that requires the platform to bridge that gap rather than leave it to chance
RISK
Pre-launch risk assessments required before new chatbot rollouts
Every new chatbot product or major update must undergo a formal risk assessment before it is made available to minors. This provision was designed to prevent the pattern that defined the social media era: companies launching products for children, discovering harms after widespread adoption, and only reluctantly adding protections under public pressure
FEEDS
Addictive social media features banned for users under 16
Infinite scroll, autoplay video, and personalized algorithmic feeds built on user history and profile are now prohibited for users under 16 on social media platforms. Platforms must either eliminate these features for minors or keep those users off their service entirely — a provision the Ninth Circuit has already upheld against legal challenge
TOYS
Four-year moratorium on AI-powered chatbot toys for children under 16
No new AI-powered chatbot toys may be sold or marketed to children under 16 in California until 2030. This provision reflects growing concern among child development researchers about AI companion devices marketed to very young children — among the least regulated AI products currently available to families
FINES
Up to $1 million per child in civil penalties for violations
The highest-stakes children's AI enforcement regime in the nation. School districts and parents also gain the right to terminate contracts with non-compliant providers and seek damages. The AFT and Microsoft's National AI Safety & Privacy Standard for schools, announced the same week, creates a parallel enforcement pathway through contracts
Sources: Governor Newsom's Office, September 10, 2026; CalMatters; Benton Institute for Broadband & Society; Wikipedia: California AI laws; AFT/Microsoft press release, September 9, 2026.
"We find ourselves in a really terrible situation in terms of what AI is doing to kids. The industry is repeating every mistake it made with social media — and the stakes are even higher this time." — Tom Siegel, founder of Google's Trust & Safety team, Reuters, September 2026
Why California's Law Protects One State — and What Every Other Parent Is Facing Right Now
Adam's Law is a landmark. It is also the law of one state. For the 88 percent of American children who do not live in California, the chatbot their child talks to late at night has no legal obligation to detect a suicidal crisis, no required parental control, and no pre-launch safety review. Whether a platform does any of these things is still entirely voluntary — left to whatever the company decides is good for business. That is the gap Tom Siegel is describing when he calls for a slowdown: not that no good work is being done, but that the companies building the products most accessible to children are moving faster than any protective framework can follow.
The protection gap — what families outside California face right now
NO LAW
No federal chatbot safety law for children exists
Senator Markey's Youth AI Privacy Act, introduced in March 2026, is still in committee. The Children CHAT Act is moving through Congress but has not passed. COPPA — the primary federal children's online privacy law — was written in 1998 and does not address AI chatbot interactions. There is no federal equivalent to California's crisis protocol requirement
PATCHWORK
134 new laws in 2026 — but most don't cover companion chatbots
Of the 134 new AI-related laws affecting children enacted or introduced across the U.S. in 2026, most address narrow issues: data collection, transparency, or school use policies. Very few address the companion chatbot scenario — an AI that builds emotional relationships with minors — and fewer still create enforcement mechanisms with real financial consequences
TOOLS
Voluntary safeguards exist — but parents must find and activate every one
Some platforms have added voluntary protections: ChatGPT's Teen Mode, parental linking, Quiet Hours, and crisis notifications (available since August 2026) are not required by law outside California — they are opt-in features that parents must find, configure, and maintain themselves, on every platform their child uses, every time a new AI tool enters the household
Sources: Senator Markey's Office, March 2026; CDT 2026 AI Legislation Update; HeyOtto!, "134 New AI Laws for Kids in 2026"; ClaimsJournal, September 23, 2026.
"Companies have immediate steps they could take to better protect young people — including basic age verification that any of them could implement tomorrow. They simply have not." — Tom Siegel, former VP of Trust & Safety, Google, September 2026
What Every Parent Must Do This Week — Whether You Live in California or Not
California's law gives its families new legal leverage. For parents everywhere else, the protections California built still describe exactly what every child needs — they just have to be built manually, family by family. Here are four actions that give your child the protections Adam's Law now legally requires companies to provide in California.
Activate crisis protocols now — don't wait for them to become law where you liveOn ChatGPT: enable Teen Mode, link your parental account, and turn on crisis notifications — these are live features as of August 2026. On every other platform your child uses, search specifically for "crisis resources" or "mental health settings" in the support documentation. If none exist, that is the platform's answer: it has not built them yet.
Audit every social media feed for the addictive features California just bannedOn every platform your child uses, check whether infinite scroll and autoplay video can be disabled in settings — many now offer this voluntarily. On TikTok, YouTube, and Instagram, screen time dashboards and feed customization settings now exist; find them and configure them. California's law bans these as defaults for under-16s — you can match that protection manually today.
Pause AI chatbot toys — the four-year California moratorium reflects a real riskCalifornia's four-year ban on AI chatbot toys for under-16s reflects the fact that child development researchers do not yet understand the long-term effects of AI companions on young children's emotional and relational development. If you have an AI-powered companion device in your home, check whether it has parental controls — and consider pausing its use until clearer safety standards exist.
Contact your state representative — and tell them what California just passedCalifornia's law is a model. Senator Markey's Youth AI Privacy Act and the Children CHAT Act are federal bills that would extend comparable protections nationally — and they need constituent pressure to move. Five minutes on your state legislature's contact page, citing Adam's Law specifically, puts the demand on the record. Thirty-four states enacted new children's tech safety laws in 2025 and 2026; the families who spoke up made it happen.
California did not pass Adam's Law because everything was fine. It passed because a teenager named Adam Raine died, and the state decided that what happened to his family should not require legal action after the fact — it should be prevented in the first place. The law is new. Companies have timelines for compliance. But the direction it points — crisis protocols, parental visibility, pre-launch risk assessment, and meaningful financial consequences — is the exact direction every family in the country needs to be demanding, from their platforms and from their legislators, right now.
You may not be in California. But you can build Adam's Law's protections into your family's digital life this week — and demand your lawmakers make them universal.
California law requires these protections. Blaick delivers them everywhere else.
Start a free 14-day Blaick trial — no credit card required.
Blaick gives your family the visibility that California's Adam's Law now legally requires — parental alerts, chatbot monitoring, addictive feature tracking, and crisis flag detection — regardless of where you live. See what every AI tool on your child's device is doing before a risk becomes a crisis. No credit card needed to start.