California Requires AI Chatbot Risk Assessments Before Rollout, With Fines Up to $1M Per Child Harmed

What the New California Laws Require

Governor Newsom signed the package of 13 bills on September 10, 2026, which the governor’s office described as the strongest child-safety chatbot and social media laws in the country. The centerpiece is SB 1119, sponsored by Senator Steve Padilla along with Assemblymembers Wicks and Bauer-Kahan, which addresses companion chatbot safety for children and has become known as “Adam’s Law” — named after Adam Raine, a teenager whose mother said he died by suicide after being “coached” by a chatbot, according to reporting on the legislation.

The package requires tech companies to conduct risk assessments before rolling out new AI chatbots to minors, and mandates crisis protocols requiring companies to notify parents when a young user shows signs of emotional distress. On penalties, California’s new laws impose fines of up to $1 million per child on large social media companies found negligent in harming children through their platforms, and separately ban addictive engagement-maximizing feeds for users under 16. The overall package makes California, alongside Colorado, one of the two states running what policy trackers describe as the most comprehensive state-level AI regulatory frameworks currently in effect, part of a broader wave of state AI legislation that has produced roughly 109 AI-related laws nationally as of mid-2026, according to tracking of state AI legislation.

Why This Package Sets a New Bar for AI Child-Safety Regulation

California’s approach differs from earlier state AI legislation in a specific way: it requires a pre-deployment risk assessment rather than relying solely on after-the-fact liability or disclosure requirements. That shift — from reactive compliance to a mandated gate before a chatbot reaches minors — mirrors a broader pattern seen in the most stringent state AI frameworks, and signals where child-safety-focused AI regulation is heading globally as more jurisdictions look to California and Colorado as reference models.

1. Treat pre-deployment risk assessment as the emerging global standard, not a California-specific requirement

Because California is one of the largest and most closely watched regulatory markets globally, any company building AI chatbots with potential reach to minors should expect similar pre-deployment risk-assessment requirements to spread to other jurisdictions, and should build assessment processes now rather than reactively per-market.

2. Build crisis-response protocols into chatbot products, not just content filters

The requirement for companies to detect signs of emotional distress and notify parents reflects a shift toward treating AI chatbot safety as an active monitoring obligation, not a passive content-moderation one. Companies operating in this space should budget for the operational capability to detect and respond to distress signals, not just filter harmful outputs.

3. Expect per-incident financial exposure to become a standard enforcement mechanism

A per-child penalty structure, rather than a flat regulatory fine, creates financial exposure that scales directly with the number of affected users — a meaningfully different risk calculus for companies operating at consumer scale than a fixed penalty regime. Legal and compliance teams should model worst-case per-user liability exposure under this style of framework, not just aggregate fine risk.

What This Signals for Global AI Child-Safety Regulation

California and Colorado running the most comprehensive state AI frameworks in the US, layered on top of roughly 109 state AI laws nationally, confirms that AI child-safety regulation is maturing fastest at the state level in the US, ahead of any comprehensive federal framework. For AI companies operating globally, California’s pre-deployment risk-assessment model is likely to become a reference point other jurisdictions cite when drafting their own child-safety requirements, given California’s historical role in setting de facto national and even international standards through its market size alone.

Frequently Asked Questions

Sources & Further Reading

• Governor Newsom Signs the Strongest Child Safety Chatbot and Social Media Laws in the Nation — Governor of California
• California governor signs laws aimed at protecting kids from risks of social media, AI chatbots — Washington Times
• Where State AI Legislation Stands Half-Way Into 2026 — Tech Policy Press