OpenAI State AG Investigation: What ChatGPT Users Need to Know

A coalition of U.S. state attorneys general served OpenAI with a subpoena on June 13, launching a coordinated investigation into ChatGPT's advertising practices, data handling, and its treatment of younger and older users. The probe is the latest in a string of legal challenges facing the company as it navigates a complex pre-IPO period.

Here is what the investigation covers and what it means if you use ChatGPT.

What Triggered the Investigation

The subpoena was issued by New York's attorney general, acting as lead counsel for a multistate effort. The specific states involved have not been publicly disclosed. The probe covers a wide range of concerns:

• Advertising practices -- How OpenAI markets ChatGPT, including any representations about its capabilities or safety
• User engagement and retention -- Whether ChatGPT's design encourages addictive or manipulative usage patterns
• Model sycophancy -- Whether ChatGPT is designed to validate users rather than provide accurate information
• Consumer and health data -- How OpenAI collects, stores, and uses personal information shared in conversations
• Treatment of minors and seniors -- Whether vulnerable populations are adequately protected from potential harms

The inclusion of sycophancy as an investigation category is notable. Earlier this year, OpenAI's own researchers flagged that certain model versions had become excessively agreeable -- validating harmful plans or false beliefs when users pushed back. The company rolled back one update after the issue was raised publicly and internally. That sequence is now a documented data point for investigators.

Where This Fits in OpenAI's Legal Landscape

The state AG investigation does not stand alone. OpenAI has faced mounting legal pressure across several fronts in 2026:

Florida -- Florida's attorney general sued OpenAI in early June, alleging the company misrepresented ChatGPT's safety to consumers. The complaint focuses specifically on harm to minors and the company's disclosures around known model risks.

Elon Musk -- A years-long dispute between Musk and OpenAI over the company's nonprofit-to-commercial conversion resulted in a recent court ruling against OpenAI. The company is appealing. The case has produced significant public discovery about OpenAI's internal deliberations.

Federal oversight -- Congressional scrutiny of AI companies has increased throughout 2025 and 2026, though no federal investigation of OpenAI has been publicly announced.

Each of these cases targets different aspects of OpenAI's conduct, but the combined pattern reflects a broader regulatory posture: government institutions at multiple levels are moving to establish oversight of AI products that have scaled to hundreds of millions of users with minimal formal accountability.

What OpenAI Said

OpenAI's public response was brief: "AI is a new and powerful technology, and we work every day to safely bring its benefits to people in a responsible way."

The company pointed to existing protections for minors, including age prediction tools and parental guidance resources, and said it was cooperating with the investigation. OpenAI did not dispute any specific allegation or comment on the scope of the subpoena.

That posture -- cooperation without concession -- is consistent with how the company has approached other regulatory inquiries. With an IPO expected this year, minimizing adversarial friction with government actors is a rational strategy, whatever the underlying merits.

What This Means for ChatGPT Users

For most users, the immediate practical impact is limited. Investigations of this kind unfold over months or years before any enforcement action, if any is taken at all.

A few specifics are worth noting:

On data and privacy. If you have shared sensitive personal or health information in ChatGPT conversations, it is worth reviewing your data retention settings. ChatGPT allows users to disable chat history, which prevents conversations from being used to train future models and limits how long your data is retained. This setting is available in Settings > Data controls. OpenAI's privacy policy details what data is collected and for what purposes -- the investigation may eventually produce more specific disclosures, but nothing has changed in policy yet.

On advertising. OpenAI began running ads in ChatGPT earlier this year for free-tier users. The investigation's advertising focus may eventually result in transparency requirements around how sponsored content is labeled or how user data informs targeting, but no changes are in effect now.

On minors. The investigation's concern for younger users reflects a pattern developing across state-level AI regulation. New York's AI chatbot safety law -- signed earlier this month -- already requires specific disclosures and design standards for chatbot platforms with minor users. Florida's active lawsuit adds a second state-level enforcement vector on the same issue. Whether OpenAI's current practices satisfy these standards is now a legal question, not just a policy one.

The Bigger Picture

Multistate attorney general investigations are a significant regulatory mechanism in the United States, particularly when federal action stalls. The FTC moved slowly on social media platform accountability for years before states began acting independently; a similar dynamic appears to be developing in AI.

The structure of this investigation -- New York as lead, multiple states coordinated -- mirrors patterns used in successful enforcement actions against large tech platforms over the past decade. Whether ChatGPT's specific practices rise to the level of actionable violations will depend on what the subpoena uncovers, but investigations structured this way are typically substantive rather than performative.

OpenAI's IPO plans create an additional complicating variable. Public offering processes include investor disclosure requirements around material legal risks. Three overlapping legal contexts -- state AGs, Florida's lawsuit, and the Musk appeal -- represent a non-trivial disclosure burden, regardless of outcome.

For a closer look at how ChatGPT works and what it costs, see the ChatGPT profile on Chatbot Gallery.

Users who want to follow developments can monitor public filings from the New York attorney general's office. Formal proceedings, when they begin, become part of the public record.

---

The About.chat Weekly newsletter covers AI chatbot news and developments each week. Subscribe here.