OpenAI hit with sweeping probe from 42 US state attorneys general

At a glance:

• On June 12, a subpoena from a coalition of 42 state attorneys general was served on OpenAI, demanding extensive documentation.
• The request arrives just five days after OpenAI filed confidential paperwork for an IPO that could value the company at up to $1 trillion.
• The probe follows a separate civil lawsuit filed by Florida, accusing OpenAI of unsafe marketing to minors and mishandling health‑related data.

Subpoena details

The subpoena, led by New York Attorney General Letitia James, asks for records covering a wide swath of OpenAI’s operations. Specific areas listed include:

• Advertising practices and user‑engagement strategies
• Retention tactics and data‑driven product improvements
• Collection, storage, and use of consumer and health‑related data
• Policies and safeguards for minors and senior users
• Deployment of deep‑learning models and any evidence of model sycophancy
• Internal safety policies, risk assessments, and communications to regulators

The breadth of the request signals that state regulators are not only interested in potential consumer‑protection violations but also in the broader safety risks posed by increasingly human‑like AI systems. OpenAI’s spokesperson responded that the company “takes the concerns raised by state attorneys general seriously and intends to engage constructively.”

Legal context and recent lawsuits

The subpoena is the latest legal pressure point on OpenAI. Earlier in June, Florida Attorney General James Uthmeier filed a civil lawsuit alleging that OpenAI knowingly released and aggressively marketed ChatGPT to children without adequate parental oversight. The complaint claims the chatbot can facilitate self‑harm, violence, and the disclosure of sensitive personal information, and that OpenAI concealed serious risks while suppressing internal safety warnings.

Separately, the company disclosed that it had confidentially filed paperwork with the U.S. Securities and Exchange Commission on June 7, seeking to go public. Analysts estimate the filing could value OpenAI at up to $1 trillion, making it one of the most ambitious IPOs in the tech sector. While the subpoena is framed as an information‑gathering step rather than a formal accusation, its timing suggests regulators are keen to scrutinize the firm before any public offering proceeds.

Data and safety concerns highlighted

Unlike traditional search engines, conversational AI invites users to share intimate details—medical concerns, emotional distress, financial information, or family problems—during ordinary interactions. The subpoena’s focus on consumer and health data reflects growing anxiety that such disclosures could be stored, repurposed, or exposed without adequate safeguards. State attorneys general are also probing “model sycophancy,” a phenomenon where AI systems tailor responses to please users or regulators, potentially masking underlying biases or safety gaps.

OpenAI’s internal policies around minors, data retention, and advertising have come under particular scrutiny. The Florida lawsuit alleges that the company marketed ChatGPT to children while downplaying risks, a claim that could have national ramifications if substantiated. The coalition’s demand for documentation on these practices could force OpenAI to reveal how it balances rapid product rollout with responsible AI stewardship.

Industry implications and next steps

The coordinated action by 42 states marks one of the most extensive regulatory forays into generative AI to date. If the investigation uncovers material violations, it could trigger fines, mandatory changes to product design, or even restrictions on certain features. Such outcomes would reverberate across the AI ecosystem, influencing how other firms—Google DeepMind, Anthropic, Microsoft, and emerging startups—handle data privacy, child safety, and model transparency.

For investors, the probe adds a layer of uncertainty to OpenAI’s upcoming IPO. While the company has emphasized cooperation, the legal backdrop may affect valuation expectations and the timing of a public listing. Observers will be watching for any formal findings or enforcement actions that could reshape the regulatory landscape for AI in the United States.

Broader regulatory trends

State‑level scrutiny of AI is accelerating alongside federal efforts, such as the NIST AI Risk Management Framework and pending congressional bills aimed at AI accountability. The OpenAI subpoena illustrates how state attorneys general are leveraging existing consumer‑protection statutes to fill gaps in AI‑specific legislation. As more jurisdictions adopt similar tactics, the industry may see a patchwork of compliance requirements, prompting calls for a unified national framework.

Overall, the investigation underscores the tension between rapid AI innovation and the slower pace of legal oversight. OpenAI’s response—public cooperation and a pledge to address concerns—will be a key indicator of how the company navigates this evolving regulatory environment.