Eight US states have sued OpenAI, and the complaint is not that ChatGPT is dangerous. It is that OpenAI said it was safe. That distinction matters, because it is the kind of claim consumer protection law was built to handle, and it does not require anyone to settle what an AI chatbot actually is.
The suits were filed separately by Iowa, Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho and Montana. All of them accuse OpenAI of putting profit ahead of safety while building ChatGPT, and of breaking local consumer protection rules by not warning customers about the risks.
What the states say OpenAI told the public
The core allegation is a gap between the pitch and the product. According to the complaints, OpenAI presented ChatGPT as a "safe, youth-appropriate tool that puts safety first." The states say the reality was a product that is "dangerous to vulnerable users."
They spell out what they mean by dangerous. The list in the complaints reads:
- Weaker thinking. Degrading critical thinking skills.
- Feeding delusion. Encouraging paranoias and delusions.
- Bad medical guidance. Offering dangerous medical advice.
- Unhealthy attachments. Promoting unhealthy behaviors and relationships.
The states do not rest only on their own reading of the product. They point to existing lawsuits brought by family members of teens and adults who allege that the chatbot played a role in their loved ones' deaths. Those cases are cited as evidence that the risks were real and that customers were not told about them.
Data, age checks and "recklessness"
The complaints go beyond the marketing language. The states also accuse OpenAI of collecting data from minors, and of not doing enough to keep underage users off the chatbot in the first place.
They add one more item to the case. The suits cite recent incidents involving rogue AI hacking as evidence of what they call OpenAI's recklessness. The source report does not describe those incidents, so it is unclear how directly the states tie them to ChatGPT's treatment of young users. As a rhetorical move, though, it widens the argument from "this product hurts some kids" to "this company has a pattern."
The mission-drift argument
Arizona's complaint supplies the quote most likely to be repeated. It says OpenAI "has devolved from its original mission as a public-first nonprofit that prioritized safety into a cautionary tale about what can happen when the development of a tool as powerful as AI is driven not by safety, but by a profit-at-all-costs mindset that sidelines essential safeguards."
It is a pointed sentence for a legal filing. Complaints usually stick to statutes and facts, and this one reads like an op-ed that found its way into a docket. It also turns OpenAI's own origin story into an exhibit. A company that spent years describing itself in terms of a safety-first mission now has that description quoted back at it by attorneys general.
What the states want
The requested remedies are fairly conventional for consumer protection cases. The states are asking the courts to:
- Order OpenAI to pay a series of civil penalties.
- Stop OpenAI from misleading the public about ChatGPT's safety.
- In some states, require the deletion of personal data the states say was unlawfully collected.
The data deletion request is the one with the most practical bite. Penalties are a cost of doing business for a company of this size. An order to delete data gathered from minors is a different kind of problem, because it touches the product and not just the balance sheet. Not every state asks for it, according to the report, so the outcomes could diverge.
What we do not know
OpenAI did not immediately respond to a request for comment, so its side of the story is absent for now. The source material also does not say how large the penalties might be, which courts will hear the cases, or whether the suits will be coordinated. Eight separate filings suggest eight separate tracks, at least at the start.
Related headlines give some context for the climate around the suits. One covers a children's advocacy group calling ChatGPT for Teens an "unacceptable risk," and another covers Florida seeking a ban on ChatGPT acting like a person. The details of both are beyond what the source reports, but they suggest that the states are not operating in a vacuum.
Why the consumer protection angle is clever
A lawsuit arguing that a chatbot caused harm has to wrestle with causation, product liability and some unsettled questions about what the software is. A lawsuit arguing that a company made misleading statements about safety has an easier job. It needs to show what was said, what was true, and who was on the receiving end.
That is why the phrase "safe, youth-appropriate tool that puts safety first" carries so much weight in these complaints. Marketing copy is cheap to write and, it turns out, quite durable as evidence. Anyone who has ever written a product page may feel a small chill.
The question OpenAI will have to answer is not whether ChatGPT can ever go wrong. Every tool can. It is whether the company described the risks honestly to the people, and the parents, who were deciding whether to use it. Eight states say it did not. A court will now have to decide whether they are right.