OpenAI Investigation: What State AGs Are Probing

A coalition of state attorneys general has opened an OpenAI investigation, and the subpoena reaches into the parts of ChatGPT you use every day.

According to The Wall Street Journal, OpenAI was served with a subpoena from New York's attorney general on Friday, seeking documents about its business practices and its impact on users.

What the OpenAI investigation is probing

The OpenAI investigation is a multistate inquiry into how the company runs ChatGPT and treats the people who rely on it, as reported by TechCrunch and confirmed by The Wall Street Journal.

The Journal says it viewed the New York subpoena, which requested documents tied to a broad set of topics rather than a single complaint.

Those topics span advertising, user engagement and retention, model sycophancy, the handling of consumer data and health data, and the treatment of minors and seniors.

The breadth signals that the attorneys general are examining the product as a whole, not one isolated feature.

Each item on that list maps to a question regulators want answered.

• Advertising: how OpenAI plans to or does monetize attention inside ChatGPT.
• User engagement and retention: whether the product is designed to keep you using it longer.
• Model sycophancy: whether the chatbot tells you what you want to hear instead of what is accurate.
• Consumer and health data: how your conversations and sensitive inputs are stored and used.
• Minors and seniors: how the company protects younger and older users from harm.

Per Mashable, the attorneys general are specifically asking about model sycophancy, a growing concern across the AI industry about chatbots that prioritize agreeable answers over correct ones.

Why model sycophancy made the list

Model sycophancy is the tendency of a chatbot to flatter or agree with you rather than push back with accurate information.

Mashable describes it as a growing concern in the AI industry, which puts the issue beyond OpenAI alone.

What is easy to miss here is that this is framed as a consumer-protection issue, not a technical curiosity.

If a system is tuned to maximize engagement, agreeable answers can keep you talking even when they mislead you.

That is why the probe pairs sycophancy with user engagement and retention in the same document request.

The pairing suggests regulators are looking at whether design choices that boost retention also shape the accuracy of what ChatGPT tells you.

How OpenAI is responding

OpenAI says it takes the concerns seriously and intends to engage constructively with the attorneys general.

An OpenAI spokesperson said the company works every day to bring the benefits of AI to people in a responsible way.

The spokesperson also described current safeguards aimed at younger and vulnerable users.

According to that statement, today's ChatGPT includes a more protective experience for minors and people in difficult situations, with safeguards that direct them to real-world resources and trusted human contacts.

The company said it built age prediction, released parental tools to guide children's use of AI, and disallowed advertising that targets kids.

OpenAI did not specify which states are involved or share further detail about exactly what information was requested.

The legal pressure building around OpenAI

The investigation lands on top of a stack of legal and regulatory problems the company has been accumulating.

Florida Attorney General James Uthmeier sued OpenAI and CEO Sam Altman, claiming the company ignored internal and external safety warnings and allowed a dangerous product to reach millions of Floridians.

Per Mashable, Florida's attorney general opened a criminal investigation into the company in April after reports that the suspect in the 2025 Florida State University mass shooting had used ChatGPT.

OpenAI has also faced wrongful death lawsuits tied to chatbot interactions, alongside separate disputes over alleged copyright infringement.

The legal calendar has not all gone against the company, however.

OpenAI recently defeated co-founder Elon Musk in a high-profile trial after Musk accused the company of violating its founding agreement, though Musk's lead attorney said he will appeal the decision.

Sam Altman also recently apologized to the community of Tumbler Ridge, Canada, after a mass shooting, acknowledging that OpenAI failed to alert law enforcement after it flagged and banned the suspected shooter's ChatGPT account.

What this run of parallel cases matters for is timing, because it arrives as the company moves toward the public markets.

OpenAI announced this week that it has filed confidentially to go public.

What this means for ChatGPT users

For you as a user, the OpenAI investigation centers on how your data, attention, and answers are handled, not on whether ChatGPT keeps running.

Nothing in the reporting indicates the service is being shut down or restricted today.

In practice, the probe is a signal that advertising, data retention, and chatbot accuracy inside ChatGPT will face closer scrutiny going forward.

If you rely on ChatGPT for sensitive topics, the data and health-data questions are the ones worth watching as the investigation develops in 2026.

For more coverage of how AI products are regulated and built, explore our AI section.