California's Child Online Safety Package: AB 2 and Adam's Law

This is the second of two articles breaking down California's child online safety package. Part 1 covered AB 1709, AB 2246, and AB 1856, addressing addictive design features, children's data practices, and age assurance. This part covers AB 2, which introduces enhanced damages for families of harmed children, and Adam's Law, which establishes protections for children interacting with AI companion chatbots.

AB 2: Enhanced damages

AB 2 allows parents to seek enhanced damages from large social media companies where negligence causing harm to a child is proven in court. Damages are set at three times the amount of the child's actual harm, with a range of $5,000 to $1,000,000 per violation.

AB 2 differs from the other three bills in the package in one significant way. It creates a private right of action. Parents can bring cases directly rather than relying on the Attorney General to act. That distinction matters because it opens a direct legal route for families affected by platform negligence, without depending on prosecutorial resources or priorities.

These provisions do not apply to cases pending on or before January 1, 2026.

Adam's Law: AI companion chatbots

Senate Bill 1119, known as Adam's Law, establishes what the California legislature describes as the most comprehensive digital protections for children interacting with AI companion chatbots in the United States. The bill passed the Senate unanimously and the Assembly 64 to 4.

The law covers several distinct areas. Platforms must verify users' ages using California's privacy-protective age-bracket signal and carry out mandatory risk assessments before releasing new or substantially modified chatbots. Under-18 users must be covered by protective defaults that only a parent can change, including no notifications, time limits on usage, and limitations on persistent conversational memory. Platforms must provide in-app referrals to mental health resources, and if an operator becomes aware of a credible threat of imminent self-harm or that a child has engaged in self-harm, they must notify a parent. Operators are liable for failure to prevent several categories of harmful output: self-harm content, sexually explicit material, romantic roleplaying, excessive praise or flattery, and emotionally manipulative content that fosters reliance or promotes isolation from friends and family. Targeted advertising to children in conversational chats is strictly limited. Independent audits of compliance must be submitted to the California Attorney General, and families have a private right of action for violations relating to crisis referral services, parental notice, default protections, and prohibitions on harmful outputs.

What the full package amounts to

Taken together, the five bills cover addictive design features, children's data and privacy defaults, age assurance infrastructure, enhanced damages for negligence, and AI chatbot regulation.

Fairpatterns detects the addictive design features and dark patterns these bills target automatically, across platforms and at scale.

‍