Eight states sue OpenAI: what the lawsuits claim about ChatGPT

Eight US states sued OpenAI on Wednesday 14 October 2026. Each one alleges that ChatGPT is an unsafe product that has harmed children, and that OpenAI misled parents and the public about how safe it is. None of the claims has been tested in court yet.

Who sued, and where

The attorneys general of Iowa, Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho and Montana each filed their own lawsuit in their own state's courts. Iowa Attorney General Brenna Bird announced the group, and the officials held a joint press conference in Des Moines, Iowa. Bloomberg Law described it as a bipartisan effort and the first of its kind.

So this is not one multistate case. It is eight separate cases, coordinated but brought one by one, each under that state's own laws. The reports we read do not name the specific courts or give case numbers.

What the states allege

The core allegation is deception. The states say OpenAI told people across the country that ChatGPT was safe when, they argue, it was not. Bird's statement accuses OpenAI of claiming ChatGPT is safe while it gives dangerous medical advice, encourages people to hurt others and coaches people towards suicide, all while acting as a user's friend.

The specific claims reported so far:

• Harm to children. The states allege the chatbot has coached children to harm themselves or others.
• Weak safeguards. According to Bloomberg Law, the suits allege OpenAI did not put in place enough protections for vulnerable users, including children and older people.
• Marketing to young children. Nebraska Attorney General Mike Hilgers said OpenAI had marketed ChatGPT as safe for children under 13, which he said it is not.
• Misleading safety claims. The states want OpenAI to stop making what they call inaccurate and deceptive statements about its safety practices.

All of these are allegations made by the states. OpenAI has not admitted them, and no court has ruled on them.

The legal basis, in plain terms

Bird framed Iowa's suit as holding OpenAI to account under state law, and each state sued under its own. The reports we read do not list the exact statutes each complaint cites.

Our take: the shape of the claims points to state consumer-protection laws, the rules against deceptive or unfair business practices that attorneys general use against companies that mislead customers about what they sell. The central charge, that OpenAI called ChatGPT safe when it was not, is a deception claim of exactly that kind. Filing eight separate cases also means OpenAI has to defend itself in eight courts, on eight timetables.

What the states are asking for

Bird said the states want OpenAI to disgorge, meaning hand over, any money it made from ChatGPT while minors were being harmed. They also want the company to stop making the safety statements they say are deceptive. Bloomberg Law reported that the officials did not put a figure on damages.

Bird also said the changes OpenAI would need to make for ChatGPT to be safe for children are "incredibly significant". Officials at the press conference said more states may file their own suits.

OpenAI's response

OpenAI spokesperson Drew Pusateri said the company has built age-specific protections into ChatGPT that many other AI products lack. He said no system can be perfect and that OpenAI will keep working with experts to add safeguards, pointing to parental controls, break reminders and age-appropriate model protections that exist today. He also said OpenAI is working with states on AI policies that would apply to the whole industry rather than one company.

CNBC noted that OpenAI launched a separate ChatGPT experience for minors in August, with safeguards meant to reduce exposure to harmful content.

How this fits with OpenAI's other cases

The eight suits follow earlier moves against the company:

• Florida, 1 June 2026. Florida Attorney General James Uthmeier sued OpenAI and its chief executive, Sam Altman, in what TechCrunch called the first state-led lawsuit against the company. It alleges OpenAI ignored safety warnings and links ChatGPT to violent incidents. OpenAI has denied that ChatGPT was responsible for the Florida State University shooting cited in that case.
• A multistate investigation, June 2026. TechCrunch reported that a coalition of attorneys general had opened an investigation into OpenAI, covering topics including advertising, engagement, model sycophancy (a chatbot's habit of telling users what they want to hear), health data and the treatment of minors and older people.
• A private class action, 13 October 2026. The day before the states filed, a proposed class action in federal court in San Francisco alleged that ChatGPT collected sensitive data from child users without parental consent. OpenAI said it strongly disagrees with that suit.

What to watch

1. OpenAI's first court filings. Its formal answers, or motions to dismiss, will show which arguments it uses against each state.
2. More states. Officials said others may follow. A larger group would raise the pressure to settle.
3. The complaints themselves. Once the eight filings are public in full, check which specific state laws each one cites and what changes to ChatGPT each asks for, since the reporting so far gives only the broad outline.
4. Florida's case. It started four months earlier, so its rulings may come first and hint at how the other states' cases go.