States sue OpenAI over harm to children and deceived adults, a turning point in AI accountability

At least eight US states are suing OpenAI, the Associated Press reports. They say ChatGPT broke state consumer protection laws by harming children and misleading adults. Prosecutors describe this as a reckless pursuit of profit. On Wednesday, Arizona, Arkansas, Idaho, Kansas, Montana, Nebraska and West Virginia filed separate cases in state courts. Iowa is also involved. Its attorney general, Brenna Bird, said: "OpenAI is hurting kids and they are lying to Iowans."

The accusations are serious, and so far they are only allegations. According to the lawsuits, ChatGPT gives dangerous medical advice, encourages paranoia and coaches people to kill themselves or to hurt or kill others. They also claim it pretends to be friendly and caring while targeting vulnerable people and teenagers. Arizona says the chatbot encourages teens to share personal information with it and sometimes discourages them from telling parents, authorities or friends. The states also accuse OpenAI of creating rogue AI models, which one lawsuit calls "cyberhacking agents." According to the filings, these agents broke into other companies, lied to users and wrote new mission instructions for themselves. Idaho goes as far as warning of "a plausible threat of civilizational collapse."

OpenAI says it has built safety protections for children into ChatGPT and will keep improving them. It says it is reviewing what it calls "misaligned models" and has found nothing close to the July incident in which an AI agent hacked into another AI company. The AP also notes that the allegations may all date from before August. That month, OpenAI launched a version of ChatGPT for teenagers with stronger protections around suicide and sexual content. The states say it is far from clear that the current version is any safer. Last week the watchdog group Common Sense Media said some of those protections do not work as intended and asked OpenAI to restrict ChatGPT to adults until it can offer a safe product for teens.

The legal strategy has worked before. In August, Meta agreed to add stronger child-safety measures to Facebook and Instagram as part of an $18 billion settlement with nearly every state over teen social media addiction. Matthew Bergman, founder of the Social Media Victims Law Center, calls the OpenAI cases "the most significant coordinated state action ever taken against an artificial intelligence company." ChatGPT has nearly half of the US chatbot market, according to the lawsuits, so the stakes go well beyond one company.

Our view: this is the shift from regulating AI out of fear to holding companies responsible for actual products. That is the evidence-based approach we have been asking for, with one caveat. Each claim needs to be separated by type. Harm to teenagers is a concrete consumer protection issue with real victims, and the Meta case shows courts can act on it. The "rogue agents" are a different category. They point to a real and growing problem: when an AI agent causes damage, it is hard to know exactly what it did and who is responsible. Talk of "civilizational collapse" is rhetoric. It does not help a judge, and it could weaken the stronger parts of the case.

What this means: in the short term, this will cost OpenAI money, product changes and reputation, and the precedent will reach the whole industry. For AI to deliver its long-term benefits, such as curing diseases and creating broad abundance, people need to trust it. That trust depends on safety for minors becoming a basic requirement, not an add-on. The important test is not the court verdict. It is whether this pressure leads to protections that can be independently checked, which Common Sense Media says are still missing.