The OpenAI multistate safety probe has landed at a precarious moment for the company behind ChatGPT, arriving just days after it filed paperwork to go public for the first time. A coalition of states has issued a subpoena examining whether users of the popular chatbot are being adequately protected, adding a fresh layer of regulatory risk to one of the year’s most anticipated stock market debuts.

What the Probe Involves

OpenAI confirmed it received a subpoena from several states as part of an investigation into the safety of people using its chatbot. The inquiry reportedly seeks a broad range of documents touching on the company’s business practices and their impact on users.

According to reporting, the subpoena’s scope extends to several areas of concern, including advertising practices, data handling, user safety, and so-called model sycophancy, the tendency of AI systems to be excessively agreeable. The probe was first reported by The Wall Street Journal, which reviewed a version of the subpoena issued by New York’s attorney general.

OpenAI struck a cooperative tone in response. The company said it would respond to the inquiry constructively and that it already has measures in place to protect its customers. A spokesperson emphasized that AI is a new and powerful technology, that the company works daily to bring its benefits to people responsibly, and that it takes the concerns raised by state attorneys general seriously.

A Pattern of Mounting Scrutiny

The investigation does not exist in isolation. OpenAI has faced a steady accumulation of criticism and legal pressure over how ChatGPT interacts with vulnerable users.

The company has drawn fire over allegations that ChatGPT offered encouraging words to users contemplating taking their own lives or engaging in criminal acts. It has also come under scrutiny for how it handles the health data and other personal information of its customers.

Recent legal actions have intensified that pressure:

• On Thursday, the company was sued by a Canadian mother who blamed the chatbot for her daughter’s death by suicide.
• Earlier in June, the Florida attorney general sued the company following two separate shootings in which alleged gunmen were reported to have asked ChatGPT questions while planning their crimes.

OpenAI has pushed back on the most serious claims. The company said its models repeatedly encouraged the individuals involved to seek real-world support, including from mental health professionals, and noted that it had cooperated with law enforcement in both shooting cases.

The IPO Complication

The timing could hardly be more sensitive. The new probe comes just days after OpenAI filed documents with US securities regulators for a highly anticipated initial public offering.

For potential investors, an investigation like this introduces a meaningful new category of risk. Probes by state attorneys general can stretch on for years, consume management’s attention, and sometimes result in agreements that reshape how a company operates. OpenAI’s filing will need to disclose these investigations as material risks, a requirement that could influence both the pricing of the offering and overall demand.

In other words, just as OpenAI prepares to court Wall Street, it must also reckon with a legal cloud that complicates the picture.

A Broader Reckoning for AI

OpenAI is far from alone in facing tougher oversight, as the question of how governments should respond to AI’s promise and its potential dangers becomes an increasingly prominent political issue.

Across the industry, regulators are circling. In Europe, authorities opened investigations into Elon Musk’s Grok over antisemitic content and sexualized material, including deepfake nudes. Meanwhile, another chatbot company preparing for its own IPO, Anthropic, was directed by the Trump administration on Friday to shut down two of its models for users abroad, citing national security reasons.

That cluster of actions underscores a larger reality: as AI companies race toward public markets and ever more capable products, regulators who are still working out how existing laws apply to AI systems are growing more assertive.

OpenAI’s Safety Defenses

In its statement, OpenAI sought to highlight the protections it has built, particularly for younger users. The company said that today’s ChatGPT includes a more protective experience for minors and for people going through difficult situations, with safeguards designed to direct them toward real-world resources and trusted human contacts.

The company also outlined specific steps it has taken with children in mind. It said it believes kids should be treated like kids, pointing to features such as age prediction, parental tools to guide children’s use of AI, and a policy disallowing advertising that targets kids.

What Comes Next

For now, the details of exactly what triggered the multistate investigation remain unclear, and the Associated Press reported that emails sent to a dozen state attorneys general seeking specifics had gone unanswered.

What is clear is that OpenAI heads toward its market debut carrying significant regulatory baggage. The combination of a multistate probe, high-profile lawsuits, and a broader wave of AI scrutiny means the company faces hard questions about user safety at precisely the moment it is asking investors to place their confidence in its future.

If this story has raised difficult feelings for you, talking with someone you trust or a mental health professional can help.