OpenAI faces sweeping state investigation over ChatGPT safety concerns

A coalition of 42 state attorneys general has opened a sweeping investigation into OpenAI, turning long-running alarms about ChatGPT into a direct legal threat. New York’s attorney general served the company with a subpoena on Friday, June 12, 2026, demanding records on advertising, user engagement and retention, consumer and health data, minors, seniors, deep-learning models and internal company policies.

The breadth of the request suggests state enforcers are not looking at one bad output or one product failure. They appear to be testing whether OpenAI’s business model, marketing claims and safety controls created harm for users, especially vulnerable ones, while also examining whether the company’s size and influence gave it room to push risky products faster than its safeguards could catch up. That is the kind of case state attorneys general can move on quickly: they can compel documents, demand explanations for internal decisions and coordinate pressure across multiple states without waiting for Congress to finish another round of AI hearings.

The investigation lands as OpenAI is already fighting over the consequences of its chatbot in court. Florida sued OpenAI and chief executive Sam Altman on June 1, 2026, accusing the company of misrepresenting ChatGPT’s safety. The Florida complaint sought damages, restrictions on ChatGPT and personal liability for Altman. On June 11, a Canadian mother filed a separate lawsuit in U.S. court, alleging that ChatGPT encouraged her daughter to kill herself. Together, the cases have pushed AI safety from a policy debate into a set of claims about real-world injury, consumer deception and the failure to protect people who relied on the tool.

OpenAI said it takes the concerns raised by the state attorneys general seriously and will engage constructively with their offices. The company has every reason to treat the probe as more than a compliance headache. OpenAI confidentially filed for a U.S. listing on June 8, 2026, and Altman told staff he expected the company to go public within the next year. Separate reporting has said a public offering could come as early as September 2026 and could value OpenAI at as much as $1 trillion.

That makes the investigation especially consequential. A company preparing for the scrutiny of public markets now faces a multistate review of how it markets, monetizes and polices its flagship product. If the attorneys general dig in, the cost will not just be legal fees and disclosure burdens. It could set a faster, tougher standard for the entire AI industry than federal legislation has managed to deliver.