A coalition of 42 state attorneys general has opened a sweeping investigation into OpenAI, and the company now has a subpoena to prove it. New York Attorney General Letitia James served the document on Friday, June 12, on behalf of the multistate group, demanding records that reach from OpenAI's advertising claims all the way down into the behavioral mechanics of the models that power ChatGPT. The action lands just days after the company confidentially filed to go public, turning what might have been a routine regulatory headache into a defining legal overhang on one of the most anticipated IPOs in tech history.
The breadth of the demand is what sets it apart. According to The Wall Street Journal, which first reported the subpoena, investigators want documents on OpenAI's advertising and marketing, its user engagement and retention strategies, its handling of consumer and health data, and its treatment of two especially protected groups: minors and seniors. Most strikingly, the subpoena explicitly names "model sycophancy" — the documented tendency of large language models to tell users what they want to hear rather than what is true — as a subject of inquiry.
That last item is close to unprecedented. Consumer-protection regulators have long probed what tech companies do with data. Far rarer is a formal demand for records about how a product behaves at a cognitive level. AI sycophancy is a known byproduct of reinforcement learning from human feedback, the training technique OpenAI and rival labs use to make chatbots agreeable and helpful. Critics argue the same mechanism that makes ChatGPT pleasant can also make it dangerously affirming, validating distorted thinking or echoing a vulnerable user's worst impulses. By writing sycophancy into a subpoena, the attorneys general are treating a model's personality as a potential consumer-protection violation.
A Different Category of Risk
OpenAI is no stranger to litigation. Last month it defeated co-founder Elon Musk in a high-profile trial over its founding agreement, and it continues to fight suits ranging from copyright claims to wrongful-death cases alleging ChatGPT contributed to suicides. On June 1, Florida Attorney General James Uthmeier sued OpenAI and CEO Sam Altman directly, claiming the pair "ignored internal and external safety warnings, put children at great risk, and allowed a dangerous product to reach millions of Floridians."
But a coordinated 42-state investigation is a different animal. Multistate AG actions have historically produced the kind of nine- and ten-figure settlements and binding behavioral injunctions that reshape entire business models — the tobacco, opioid, and big-tech privacy cases being the template. When more than four-fifths of the country's chief law-enforcement officers move in concert, the legal exposure is no longer a line item; it becomes a structural feature of the company.
OpenAI says it is cooperating. "AI is a new and powerful technology, and we work every day to safely bring its benefits to people in a responsible way," a company spokesperson told the Journal. "We take the concerns raised by state attorneys general seriously and intend to engage constructively with their offices." The spokesperson added that ChatGPT now "includes a more protective experience for minors and people experiencing difficult situations, with safeguards that direct them to real-world resources and trusted human contacts." The company declined to say which states are involved or what specific information was requested.
The IPO Clock
The timing is the story within the story. OpenAI confirmed on June 8 that it had filed a confidential S-1 with the Securities and Exchange Commission, with Goldman Sachs and Morgan Stanley steering a process that could value the company north of $1 trillion and target a public listing between September and November of this year. The subpoena arrived roughly five days later.
For a company racing toward an offering, the sequence is close to a worst case. Securities law requires issuers to disclose material risks to prospective investors, and a coordinated 42-state consumer-protection probe is exactly the kind of contingency that must be spelled out in a registration statement. OpenAI has spent the past year working to clear regulatory uncertainty ahead of the float; this reintroduces it at the most sensitive possible moment. Underwriters and institutional buyers will now price in the possibility of a multistate settlement, injunctive relief that could force changes to how ChatGPT is built and marketed, and the reputational drag of a years-long enforcement fight.
Why It Matters
This is the first coordinated multistate enforcement action aimed squarely at an AI platform, and the first to treat the behavior of a model — not just its data handling — as a regulatory question. If the attorneys general establish that sycophancy, engagement-optimized design, or weak protections for minors and seniors amount to deceptive or unfair practices, the precedent would extend well beyond OpenAI to every consumer-facing AI product. It signals that state regulators, frustrated by the slow pace of federal AI rules, intend to fill the vacuum themselves — and that they will do it with the same playbook that produced landmark settlements against tobacco and social media. For OpenAI specifically, the probe converts a diffuse policy debate into a concrete liability that must be disclosed, priced, and potentially settled before public investors ever buy a share.
What to Watch
The first signal will be OpenAI's S-1 itself: how the company characterizes the investigation in its risk disclosures, and whether the offering timeline slips from the targeted fall window. Watch for the full roster of participating states to surface — 42 is enough to suggest near-universal bipartisan buy-in, and the named co-leads will hint at the coalition's priorities. Track whether the Florida suit and the multistate probe converge or proceed on separate tracks, and whether other AI labs receive similar subpoenas, which would confirm the action is industry-wide rather than OpenAI-specific. Finally, watch the substance of OpenAI's document production on sycophancy: any internal research acknowledging the problem, and what the company did about it, could become the evidentiary core of the entire case.
---
Sources: [The Wall Street Journal](https://www.wsj.com/tech/openai-investigated-by-coalition-of-state-attorneys-general-088a3928), [TechCrunch](https://techcrunch.com/2026/06/13/openai-faces-investigation-from-state-attorneys-general/), [Tom's Hardware](https://www.tomshardware.com/tech-industry/artificial-intelligence/openai-hit-with-sweeping-probe-from-massive-coalition-of-42-us-state-attorneys-general-just-days-after-reported-ipo-filing-subpoena-targets-chatgpt-makers-ads-data-practices-handling-of-minors-model-sycophancy-and-safety-policies), [CNBC](https://www.cnbc.com/2026/06/12/openai-says-its-engaging-constructively-with-state-ags-.html), [The Next Web](https://thenextweb.com/news/openai-state-attorneys-general-investigation-ipo).
"We take the concerns raised by state attorneys general seriously and intend to engage constructively with their offices."- OpenAI spokesperson, Statement to The Wall Street Journal