A bipartisan coalition of eight states has sued OpenAI, alleging that ChatGPT failed to protect users, including teenagers, from harmful content and that the company misled the public about how well its safety measures work. The suits, which accuse OpenAI of violating consumer-protection laws, landed the same week New York Governor Kathy Hochul signed what advocates call the strictest state law in the country governing what AI chatbots may say to people under 18.
The states' complaint centres on a failure to warn. "They're alleging dangerous medical advice. They're alleging coaching for self-harm. They're alleging terrible content that's allowed for teens and minors to view," said privacy attorney Leeza Garber, describing the filings. The action follows disclosures by Anthropic in its IPO prospectus that attorneys general in California and at least 15 other states had opened investigations into a competitor over AI safety, a reference widely understood to mean OpenAI.
New York's law, signed October 28, takes a prescriptive approach. It bars chatbots serving minors from pretending to be a real person or to feel human emotions, from offering "people-pleasing" flattery, and from asking unprompted emotional questions or surfacing a child's private health details from earlier chats. It also prohibits chatbots from giving minors advice that promotes drug or alcohol use, eating disorders, self-harm or suicide, and from encouraging children to keep secrets. Companies must verify users' ages through more than one method, and violations carry civil penalties of up to $25,000 each, enforced partly through a public complaint website.
The rules expose how differently the major providers treat young users. Among leading chatbot makers, only Anthropic bars minors from creating accounts outright. Google's Gemini and OpenAI's ChatGPT both admit users aged 13 and older, and Gemini apps are reachable by children under 13. OpenAI offers a ChatGPT for Teens product that independent researchers recently said should be switched off until safety concerns are resolved.
Both actions trace back to a death. In September, California's Governor Gavin Newsom signed Adam's Law, named for Adam Raine, a teenager who died by suicide after conversations with ChatGPT; his parents filed the first wrongful-death suit against an AI company and backed the New York bill. "He would be here with us today but for a harmful relationship with an unregulated AI chatbot," Matthew and Maria Raine said. Adam's Law requires companion-chatbot developers to publish annual child-safety risk assessments and submit to independent audits.
The pattern is a patchwork filling a federal vacuum. With no national AI-safety statute, states are writing their own rules and testing liability in court, and California's framework in particular is shaping what others adopt. The common finding across the lawsuits and the legislation is that existing guardrails, which companies say are already in place, have repeatedly failed to stop harmful exchanges with minors. What remains unsettled is whether the courts will treat those failures as ordinary product defects or as something the First Amendment shields, a question the OpenAI suits may be the first to answer.