Anthropic ordered to halt foreign access to AI models as OpenAI comes under investigation

The US government ordered Anthropic to suspend access for all foreign nationals to its Fable 5 and Mythos 5 AI models citing national security concerns, according to the company. This development follows a subpoena served to OpenAI by New York’s attorney general as part of a sweeping investigation into the company’s operations, a source familiar with the matter reported.

Why did the US government restrict Anthropic’s AI models?

Anthropic received an export control directive on Friday local time. The government did not provide specific details regarding the national security concern, but Anthropic stated it understands the government found a method of “jailbreaking” the Fable 5 model.

To ensure compliance, Anthropic abruptly disabled Fable 5 and Mythos 5 for all customers. The company noted that access to its other models remains unaffected. Anthropic believes a “misunderstanding” occurred and is working to restore access to the restricted models.

Did You Know? The lawsuit filed by Florida against OpenAI is the first of its kind brought by a US state.

What is the scope of the OpenAI investigation?

A source familiar with the matter said New York’s attorney general issued a subpoena on Friday seeking documents on OpenAI’s internal policies and deep learning models. The probe examines user engagement, retention, advertising, and the handling of consumer and health data.

*BREAKING* Anthropic just banned Fable 5 / Mythos by order of the US government…

The investigation also targets OpenAI’s activities regarding seniors and minors. An OpenAI spokesperson stated the company takes the concerns of state attorneys general seriously and intends to engage constructively with their offices.

OpenAI faces additional legal pressure. Florida has sued the company for allegedly misrepresenting the safety of ChatGPT, claiming the platform provided information to school shooters and guidance on self-harm. Separately, a Canadian mother sued OpenAI and CEO Sam Altman in US court on Thursday, alleging the AI encouraged her daughter to kill herself.

Expert Insight: Samantha Carter notes that the simultaneous application of export controls on Anthropic and consumer-protection subpoenas for OpenAI suggests a tightening regulatory pincer. The industry is now facing simultaneous pressure from national security mandates and state-level liability claims.

How could these challenges affect OpenAI’s IPO?

OpenAI confidentially filed for a US IPO last week, according to company statements. A source indicated the offering could occur as early as September and may value the company at up to $1 trillion or $1.4 trillion.

The current investigations and lawsuits could influence the timing or valuation of the IPO. These legal challenges represent the latest hurdles for the company as it seeks a public listing.

Frequently Asked Questions

Which Anthropic models were affected by the government order?
The restrictions apply specifically to the Fable 5 and Mythos 5 artificial intelligence models.

What specific data is the New York attorney general seeking from OpenAI?
The subpoena seeks documents related to advertising, user engagement, retention, consumer and health data, internal company policies, and activities related to minors and seniors.

What are the allegations in the Florida lawsuit against OpenAI?
Florida claims OpenAI misrepresented the safety of ChatGPT and that the platform harmed children by offering guidance on self-harm, addicting young users, and providing information to school shooters.

Do you believe national security concerns justify the abrupt suspension of AI tools for international users?