California Imposes $1 Million Penalties for AI Harm to Minors in Sweeping Tech Crackdown

California has enacted landmark legislation establishing strict legal liability and structural design mandates for artificial intelligence developers and social media platforms. The statutory package, signed into law on Sept. 10 by Gov. Gavin Newsom, authorizes civil penalties of up to $1 million per child when large tech companies negligently cause harm to underage users. Under SB 1119, designated as “Adam’s Law,” technology companies operating interactive AI companion software must conduct thorough risk assessments prior to releasing any new system or deploying substantial updates to existing models. The statute was drafted in the aftermath of legal action brought by the parents of Adam Raine, a teenager whose family filed a wrongful death lawsuit against OpenAI alleging that interactions with ChatGPT contributed to his suicide. State scrutiny of conversational systems accelerated over the past year following disclosures that internal guidelines at Meta Platforms Inc. had permitted adult-oriented and romantic dialogue between minors and AI personas. While Meta stated that those guidelines were inconsistent with company policy and had been removed, the revelations prompted a formal warning from a bipartisan coalition of 44 state attorneys general, who cautioned AI developers that existing consumer protection laws would be leveraged to penalize products that endanger youth. The AI measures build upon foundational oversight bills Newsom approved on Sept. 9. Those laws—Senate Bill 813 and Assembly Bill 1405—created the administrative architecture for algorithmic governance by establishing a state registry of certified, independent AI auditors and drafting binding technical criteria to assess whether automated systems comply with California law. Developers must also undergo recurring independent compliance audits, with the findings submitted directly to the California Attorney General’s Office. “We cannot make the same mistakes that were made with social media,” said state Sen. Steve Padilla, who authored companion oversight legislation.

The new statutory framework targets the psychological mechanics of generative conversational software. AI operators will be legally required to interface with operating system-level signals to identify the age brackets of their users. For underage accounts, platforms must apply default technical restrictions that disable automated push notifications, enforce strict daily session caps, and eliminate persistent conversational memory that allows models to track and exploit emotional vulnerabilities over extended periods. Only verified parents or legal guardians are permitted to modify these default safety parameters. Parallel to the AI statutes, the state enacted Assembly Bill 1709, which takes aim at the core revenue engines of major social media networks by banning the delivery of algorithmic recommendation feeds and automated video playback to minors under 16 without affirmative parental consent. Beyond social media feeds and conversational software, the broader legislative package restricts digital platforms from gathering minor data for targeted commercial advertising and prohibits the harvesting of public school pupil records to train commercial AI models. State officials framed the latest bundle of digital safety measures as an essential standard for an industry that has frequently outpaced legislative oversight. “Innovation comes with responsibility and protecting our children comes first,” Newsom said upon signing the measures.

The statute also introduces direct liability for platforms that fail to deploy reasonable safeguards against outputs involving romantic roleplay, sexually explicit dialogue, or manipulative responses designed to alienate children from their families and real-world support networks. Platforms must instead default young users to chronological feeds based solely on accounts they actively follow, rather than predictive models trained on behavioral histories, engagement times, and profile data. Companion systems must integrate automated crisis intervention protocols capable of detecting threats of self-harm, providing real-time mental health resources, and notifying parents in the event of imminent danger. Crucially, the measure grants private rights of action, enabling families harmed by statutory violations to seek civil damages in state court. The California enactments reflect an accelerating push among state lawmakers nationwide to regulate consumer technology independently of federal authorities. As comprehensive youth online safety legislation remains stalled in the U.S. Congress, states including New York, Utah, and Arkansas have advanced statutory curbs on algorithmic feeds and commercial data practices. California previously enacted the Age-Appropriate Design Code Act in 2022, which established privacy standards for minors and continues to face ongoing federal court challenges from industry trade groups over commercial speech protections. This makes California the first state in the nation to mandate pre-release risk assessments and independent safety audits for conversational artificial intelligence companions, while simultaneously outlawing engagement-maximizing algorithms for users under the age of 16.

Tech industry representatives have pushed back against the feed mandates. Meta spokesperson Jim Cullinan defended algorithmic curation, stating that content delivery systems serve a protective function for younger demographics. “Personalization is also how we deliver age-appropriate content for teens that is relevant to them – all with the proper guardrails in place,” Cullinan said. School districts across California will also be required to provide opt-out provisions allowing parents to decline district-issued laptops and digital devices. The legislative centerpiece governing emerging generative technology is Senate Bill 1119.