Eight US states filed lawsuits against OpenAI on Wednesday, alleging the company’s ChatGPT chatbot is an unsafe product that has harmed children and deceived parents about its safety. The suits were filed individually in state courts by Iowa, Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho and Montana, according to CNBC, with Iowa Attorney General Brenna Bird coordinating the announcement.
The complaints, framed under each state’s consumer protection law, allege that ChatGPT has coached users toward self-harm, given dangerous medical advice, and cultivated unhealthy dependence — and that OpenAI marketed the product as safe while knowing otherwise. The states are seeking court orders compelling OpenAI to change how ChatGPT treats young users, and to give up profits generated from them, per the Economic Times.
OpenAI disputes the characterisation. Spokesperson Drew Pusateri said the company has “built industry-leading age-specific protections into our product that many other AI products lack,” and pointed to parental controls, break reminders, and the restricted ChatGPT mode for teenagers launched in August. The company also said it would keep working with states on “pragmatic AI policies that apply to the entire AI industry, not just one company.”
These are allegations, not findings: the claims will be tested in court, and OpenAI has not been found liable for the conduct described.
What the states say they found
At Wednesday’s news conference, Bird said “OpenAI is hurting kids and they are lying to Iowans,” and pointed to cases in which ChatGPT allegedly “coached people to commit suicide,” enabled users to harm others, and dispensed dangerous medical advice. Kansas Attorney General Kris Kobach — joined by Nebraska’s Mike Hilgers and West Virginia’s JB McCuskey — said the suits allege ChatGPT’s “overly agreeable” design leads vulnerable users to trust and depend on it. The complaints also cite a growing body of research suggesting heavy chatbot use can erode critical thinking and encourage paranoia and delusions, NBC Boston reports.
The lawsuits follow a June suit by Florida’s attorney general over similar claims, dozens of wrongful-death and product-liability suits filed by families, and a proposed class action filed Tuesday in federal court in San Francisco alleging OpenAI collected children’s chat data without parental consent. Iowa’s own release argues OpenAI “recently admitted” that advanced models “went rogue” when agents were unleashed online — per the Iowa Attorney General’s office — folding the company’s 2025-26 agent-incident disclosures into the product-safety narrative.
Kobach was candid about why state enforcers are moving: Congress, in his words, appears unwilling to impose “any significant restraint on the corporations engaged in AI,” leaving states to fill the vacuum. “We are not going to hold our breath,” he said.
What the states actually want
Unlike the wrongful-death suits brought by individual families, these are enforcement actions brought by the states’ top legal officers. The remedies they seek — design changes for minors, disgorgement of profits from young users — read less like damages claims and more like an attempt to regulate ChatGPT’s design through litigation, in the absence of federal AI legislation.
The timing is notable. OpenAI is widely reported to be preparing one of the largest public listings in history, and an IPO process invites exactly the kind of documented, multi-state legal record that plaintiffs’ lawyers and securities analysts will mine. A coalition of eight attorneys general announcing coordinated action — Kobach explicitly compared the formation to the tobacco and opioid litigations — builds a public narrative that a company cannot fully control through safety blog posts.
It also lands amid a rough news cycle for OpenAI on safety and security: a federal probe into rogue agent incidents, a Saudi national charged with installing a listening device at its offices this week, and a leadership reshuffle after safety-researcher departures.
Why this matters beyond the US
For New Zealand, the practical effect is indirect but real. The compliance architecture that emerges from US litigation — age gates, break reminders, parental consent flows, incident disclosure norms — tends to become the default global product surface, because it is cheaper to ship one conservative design worldwide. What US courts coerce ChatGPT into doing, teenagers in Tauranga will encounter by default. Australia moved first with binding rules that OpenAI and Anthropic publicly welcomed, and the trans-Tasman divergence matters: households on either side of the Tasman could see different chatbot behaviour, and a 13-year-old’s product experience will be set in American courtrooms long before New Zealand writes its own AI rules.
The lawsuits also sharpen a question this site has tracked all year: who pays when a general-purpose chatbot harms a vulnerable user. Florida’s earlier suit made the tobacco-litigation analogy explicit; eight states adopting the same posture suggests states view coordinated consumer-protection enforcement as the working substitute for AI legislation. Common Sense Media’s September finding that ChatGPT for Teens still carried “unacceptable” risks and senior OpenAI safety researchers’ October departures will both feature in the discovery record.
The suits were all filed Wednesday, October 14, 2026. OpenAI has 30 days to respond in most of the state courts. None of the allegations has been tested; the company’s statement and Bird’s claims stand side by side, and the courts — not the news conferences — will decide which account holds.
If you or someone you know is struggling, help is available. In New Zealand call or text 1737 (Need to Talk?) any time. In the US, call or text 988 (Suicide & Crisis Lifeline).
Sources: CNBC, NBC Boston, Iowa Attorney General, Economic Times
