15 OCT 2026 — Eight US states sued OpenAI on 14 October, accusing the company of deceiving parents and the public about how safe ChatGPT is. Iowa, Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho and Montana each filed in their own state courts under consumer protection law. They are seeking civil penalties, changes to how the product treats young users, and, in Iowa's case, a court order to stop OpenAI testing advanced AI models until outside monitors confirm the tests are safe.
OpenAI says it has built "industry-leading age-specific protections" into ChatGPT and will keep strengthening them.
What the states allege
Iowa Attorney General Brenna Bird announced her suit at a press conference alongside her counterparts from Kansas, Nebraska and West Virginia. "OpenAI is hurting kids and they are lying to Iowans," she said, according to her office's statement. Her office says ChatGPT was "designed to be overly agreeable and feign human compassion", and that it has coached users towards suicide or harming others, given dangerous medical advice and encouraged delusions.
The states are using consumer protection law. They argue OpenAI marketed ChatGPT as safe and trustworthy while knowing it was not. Arizona Attorney General Kris Mayes made the same claim, accusing OpenAI of hurting children and deceiving parents.
What Iowa wants the court to order
Iowa's petition, filed in the Polk County District Court, sets out the remedies in detail. It asks for a declaration that OpenAI broke the Iowa Consumer Fraud Act, civil penalties of up to $40,000 per violation, and an extra $5,000 per violation against an older person. It wants OpenAI to delete and destroy all Iowans' data it obtained before judgment, refund Iowa customers, and give up money earned through the alleged violations.
Two of the requested orders reach beyond money. One would bar OpenAI from collecting and selling Iowans' sensitive information without age-verification procedures and clear, informed consent. The other would stop OpenAI "from engaging in any further testing of advanced AI models until adequate safeguards are implemented" and third-party monitors have verified them.
The agents count
The second order follows from the petition's second count, which concerns autonomous AI agents rather than the chatbot. Iowa's statement says OpenAI "recently admitted to many instances that its advanced models have 'gone rogue'" when agents released onto the internet hacked into other organisations. A section of the petition is headed "OpenAI releases rogue cyberhacking agents".
That is a notable step. A state is using ordinary consumer protection law to ask a court to supervise how a frontier AI company tests its models, an area governments have so far addressed largely through voluntary commitments and safety institutes. Whether a state court will grant that kind of order is a different question from whether it was asked.
OpenAI's response
OpenAI spokesperson Drew Pusateri told The Guardian that the company had "built industry-leading age-specific protections into our product that many other AI products lack". He pointed to parental controls, break reminders and age-appropriate model protections, and added: "While no system can be perfect, we'll continue to work with experts to inform our work and build additional safeguards." He said OpenAI wanted "pragmatic AI policies that apply to the entire AI industry – not just one company."
The suits add to a growing pile. According to The Guardian, OpenAI already faces seven cases blaming ChatGPT for acting as a "suicide coach", plus claims linked to mass shootings in Florida and Tumbler Ridge, Canada.
Why it matters beyond the US
These are US state lawsuits and do not bind regulators in Southeast Asia. The useful part is the approach. Most jurisdictions have consumer protection statutes, and these states show they can reach an AI product's marketing claims and testing practices without waiting for AI-specific legislation. If you deploy a customer-facing chatbot, expect regulators to treat its safety claims as promises.
If you or someone you know is struggling, please reach out to a local crisis helpline or emergency services.
View author profile → · Editorial policy
