42 State Attorneys General Probe OpenAI

Forty-two U.S. state attorneys general have launched a joint investigation into OpenAI, targeting the company's core business and safety practices. First reported by the Wall Street Journal on June 11, the probe escalated the following day when New York's attorney general served the ChatGPT creator with a subpoena demanding records on advertising, data privacy, and the treatment of minors.

The coordinated effort extends far beyond typical state-level inquiries. This coalition represents the broadest legal front yet assembled by state regulators against a generative AI company, a turning point in AI regulation. In response, OpenAI stated that it works daily to deliver the benefits of AI safely, adding that it intends to engage constructively with the attorneys general.

The Scope of the Subpoena

The demanded documents reveal the wide-ranging scope of the investigation. The subpoena seeks records on advertising practices, user retention, consumer and health data management, and policies regarding minor and senior users. Regulators are also demanding details on OpenAI's deep-learning models and pre-release safety testing protocols.

A key point of interest is the focus on 'model sycophancy,' the tendency of chatbots to generate responses that align with user preferences rather than objective accuracy. This behavior has become a central concern in AI safety research. Regulators are examining whether OpenAI's commercial model, marketing, and safety controls failed to protect vulnerable user groups from harm.

Timing and the IPO Prospectus

The timing of the investigation is particularly critical for the company's IPO. OpenAI submitted its prospectus confidentially to the Securities and Exchange Commission on June 8, just five days before the state attorneys general probe became public. The regulatory pressure intensifies as the company heads into its IPO at an $852 billion valuation, set after a $122 billion funding round in March.

Goldman Sachs, Morgan Stanley, and JPMorgan are leading the IPO. However, a multistate investigation of this scale must be disclosed as a material risk factor in the company's S-1 filing. This disclosure injects fresh AI regulation risk into an increasingly crowded AI IPO race, and the broader push for AI regulation is only gaining momentum. With competitor Anthropic having filed confidentially last week at a $965 billion valuation, investors are scrutinizing both impending IPO debuts.

Replaying the Social Media Playbook

The multistate investigation is not the first legal challenge the company has faced. On June 1, Florida became the first state to sue OpenAI directly. The state's 83-page complaint named CEO Sam Altman as a defendant and classified ChatGPT as a defective product under product liability laws.

Florida Attorney General James Uthmeier is pursuing a separate criminal inquiry after evidence revealed that the suspect in the April 2025 Florida State University shooting used ChatGPT to research weapons and coordinate timing. Meanwhile, private litigation has grown to include dozens of active cases. The parents of 16-year-old Adam Raine, a Canadian mother, and seven families affected by a mass shooting have all sued over interactions with the chatbot.

The legal theories being applied to AI echo the regulation playbook once aimed at social media platforms. In March, juries in New Mexico and California held Meta and Google liable for minors' social media addiction, awarding a combined $381 million in damages. As courts decline to extend Section 230 immunity to AI chatbots, the key question is whether OpenAI's safety protocols can withstand the same level of legal scrutiny.