US Commerce Secretary Lutnick has expressed concerns in a letter to Anthropic regarding the potential diversion of the latest Mythos and Fable models to foreign military intelligence users. This warning comes in the context of ongoing national security reviews, where U.S. authorities have mandated AI companies to implement export controls and restrict access to advanced AI models to mitigate such risks. The government has previously issued letters to AI developers highlighting these concerns, which has led to immediate compliance measures, including suspensions of access to certain models.

Anthropic: Anthropic develops frontier AI models and has recently engaged with U.S. authorities on safety and access protocols for its latest offerings. The company has suspended access to certain high-capability models following government directives aimed at compliance with export controls. This action stems directly from concerns raised in the letter from Commerce Secretary Lutnick about potential misuse by foreign entities.
US Commerce Secretary Lutnick: Howard Lutnick serves as U.S. Commerce Secretary in the Trump administration, overseeing trade policies, tariffs, and negotiations with foreign nations and domestic companies. He has emphasized national security in regulatory decisions and maintained a hands-on approach to pressuring partners on economic matters. His recent letter to Anthropic addresses risks of advanced AI model diversion.

AI Governance: Frontier AI labs are coordinating with federal officials on export controls for the most capable systems amid ongoing national security reviews.
National Security: U.S. authorities have directed AI companies to restrict access to advanced models to prevent potential use by foreign military intelligence users.
Regulatory Scrutiny: Government letters to AI developers have flagged risks of model diversion, prompting immediate compliance measures such as access suspensions.