OpenAI suddenly faces joint investigation, Anthropic’s top model halted—U.S. AI regulatory crackdown escalates

OpenAI suddenly faces joint investigation, Anthropic’s top model halted—U.S. AI regulatory crackdown escalates

Attorneys general from multiple U.S. states have jointly launched an investigation into OpenAI, putting this AI giant—currently racing toward an IPO—under a fresh wave of legal pressure.

According to Reuters citing informed sources, the New York attorney general, representing a coalition of state attorneys general, delivered a subpoena to OpenAI on Friday, demanding the company provide documents covering business areas such as advertising, user engagement and retention, consumer and health data processing. The investigation has not been publicly announced. An OpenAI spokesperson stated that the company is cooperating with the probe and will "communicate constructively" with the states' attorney general offices.

The scope of this investigation is substantial: the subpoena extends into activities related to minors and seniors, deep learning models, and internal company policies. The timing is especially sensitive for a company that just secretly filed for an IPO this week, with a possible valuation of up to $1 trillion.

Reuters, citing sources, reports that the subpoena sent by the New York attorney general on behalf of the multi-state alliance covers a wide range. In addition to advertising and user data processing, the subpoena specifically demands disclosures about activities involving minors and seniors, the workings of deep learning models, and company policy.

An OpenAI spokesperson replied: "Artificial intelligence is a new and powerful technology, and we strive every day to bring its benefits to users safely and responsibly. We take the concerns raised by the states' attorneys general seriously and will work constructively with their offices." The spokesperson further added that the current version of ChatGPT offers a more protective experience for minors and vulnerable users, along with mechanisms guiding them to real-world resources and trusted contacts.

This multi-state investigation is not an isolated event for OpenAI, but the latest development in a series of recent legal challenges. Bloomberg reported that earlier this month, Florida filed a lawsuit against OpenAI and CEO Sam Altman, accusing the company of ignoring safety warnings and launching ChatGPT despite knowing it could be harmful to users—the first such case filed by a U.S. state government.

Meanwhile, Reuters reported that a Canadian mother filed a lawsuit against OpenAI and Sam Altman in a U.S. court on Thursday, claiming ChatGPT encouraged her daughter to commit suicide. Previously, Florida’s lawsuit also accused the platform of providing information to school shooting perpetrators, offering guidance on self-harm, and causing young users to become dependent.

Multiple lawsuits are focusing on the same core issue: the potential harm of ChatGPT to vulnerable groups, especially minors.

These accumulating legal pressures coincide with OpenAI’s key moment in its sprint to the capital markets. Reuters reports, citing insiders, that OpenAI announced on Monday it has secretly filed for an IPO, which could happen as early as September this year, with the company valued up to $1 trillion. Bloomberg reports that in its March funding round, OpenAI was valued at $852 billion, making it one of the highest-valued startups in the world.

The launch of a joint multi-state attorney general investigation means OpenAI now faces legal scrutiny from the state level in addition to federal regulatory pressure.

Anthropic and OpenAI both face regulatory action—what does that signal?

As OpenAI is under regulatory investigation, its archrival Anthropic’s two most advanced models, Fable 5 and Mythos 5, have been subjected to comprehensive access restrictions by the U.S. government, citing jailbreaking security risks.

Commerce Secretary Howard Lutnick sent a letter to Anthropic CEO Dario Amodei, placing Fable 5 and Mythos 5 under export control, restricting access by all institutions and individuals outside the U.S. and all foreign nationals within the U.S., including Anthropic's foreign employees.

Anthropic has closed Fable 5 and Mythos 5 access for all users and apologized to affected customers. Amazon Web Services subsequently announced that Anthropic requested the removal of user access to Claude Fable 5 and Claude Mythos 5 across all regions. All other Anthropic models are unaffected.

Both Anthropic and OpenAI facing regulatory action points to changes in the conflict structure within the AI industry: from early internal governance disputes and founder conflicts at OpenAI—a "company soap opera"—to open competition between tech giants, now further escalating into systematic game-playing between enterprises and regulators.

Renowned AI scholar Gary Marcus joked on social media that this series of events "feels like a Netflix-produced OpenAI reality show," commenting:

“Season 1 was Sam Altman being fired and rehired; Season 2 was Elon Musk versus Sam Altman; now, this season’s plot turns to Anthropic and the U.S. government.”

He added, “I didn’t expect the storyline to get so intense so quickly.”

Notably, recent media reports indicate the U.S. government is exploring the possibility of directly holding stakes in major AI companies, and considering distributing returns as dividends to all American households.

It is reported that senior U.S. officials have held preliminary discussions with several major AI companies about government equity acquisition. The core proposal is for companies to voluntarily transfer shares to the government, rather than through coercive measures.

Both OpenAI and Anthropic are preparing sizeable IPOs—Reuters previously reported that OpenAI plans to file for a secret IPO, while Anthropic has already submitted listing documents to U.S. securities regulators this week. Analysts believe that if the U.S. government acquires shares before or after the IPO, the potential financial and policy impact would be significant.

Risk Warning and DisclaimerMarkets carry risks, investments should be made cautiously. This article does not constitute personal investment advice and has not considered the specific investment objectives, financial situations, or needs of individual users. Users should consider whether any opinions, viewpoints, or conclusions in this article are suitable for their particular situation. Investing based on this article is at your own risk.