U.S. Cites China-Russia Military Intelligence Risk, Restricts Anthropic's New AI Models Under Export Control Law

Central Message

(Translated by Central News Agency, Washington, June 15) U.S. Secretary of Commerce Howard Lutnick has directed Anthropic to restrict access to its latest artificial intelligence (AI) models, citing concerns that they could be exploited by military intelligence personnel in China and Russia. According to experts, this is the first time the Commerce Department has exercised this authority.

A letter obtained by Reuters shows that Lutnick, in correspondence with Anthropic CEO Dario Amodei, cited these risks and ordered the immediate cessation of exporting or transferring the two new AI models—Fable 5 and Mythos 5—to foreign nationals anywhere in the world.

A Trump administration official said senior technical executives from Anthropic met with Commerce Department officials today to discuss solutions.

Due to significant national security implications, the government insists on ensuring these models cannot be used to threaten U.S. interests, while Anthropic is pushing to restore access to its top-tier models. As of June 12, both Mythos and Fable top-tier models have been fully taken offline.

Insiders reveal that Lutnick remains deeply involved and regularly communicates with Amodei. Both are scheduled to attend the G7 summit in France, where they may discuss the matter in person.

Sean Cairncross, National Cybersecurity Director, also attended a working-level meeting held at the Commerce Department today.

Anthropic previously stated that the government believes there is a method to bypass safety safeguards—or 'jailbreak'—Fable 5 to discover software vulnerabilities. The company added that such vulnerabilities could only identify 'minor' security flaws, and similar issues can be found using other publicly available models.

Relations between the Trump administration and Anthropic soured earlier this year when the company refused to allow the U.S. military to use its AI models for domestic surveillance and fully autonomous weapons systems. The administration subsequently placed Anthropic on a national security blacklist.

The San Francisco-based AI startup has secretly filed for an initial public offering (IPO). Sources indicate that Anthropic had collaborated with the government to test Fable 5 prior to release and had received government deployment approval.

The Commerce Department's letter cites authority granted under the 2018 Export Control Reform Act to regulate emerging technologies critical to U.S. national security. Export control experts say this is the first time the department has invoked this power.

The letter states that exports (or transfers to foreign nationals in the U.S.) of these AI models will require licensing review and warns of 'immediate criminal and civil penalties' for non-compliance.

However, export control experts note that AI models are typically deployed via remote access rather than physical export, and current regulations do not fully cover such mechanisms. This has raised legal questions about the Commerce Department's authority to take such actions.

Over 80 cybersecurity executives and experts, including leaders from major firms like Nvidia and Adobe, signed a joint open letter yesterday supporting Anthropic's position, urging the Trump administration to lift restrictions on the company. (Translation: Hsu Hsiang-ping) 1150616