Anthropic Halts Access to Advanced AI Models Under U.S. National Security Order

Artificial intelligence firm Anthropic announced it is suspending public access to its most advanced AI models, Fable 5 and Mythos 5, in compliance with a U.S. national security export control directive. The order, originating from the Trump administration, mandates that access to these cutting-edge AI services be blocked for foreign nationals, both outside and within the United States.

Anthropic stated that the government has not provided specific details explaining why its new AI services pose a national security risk. The company issued an apology to its paying customers for the sudden interruption, emphasizing its belief that government oversight of unsafe AI developments should be transparent, fair, clear, and based on technical facts. Anthropic indicated that this directive does not adhere to those principles.

The suspension comes shortly after Anthropic made Fable 5 available to subscribers of its Claude service. Fable 5 was presented as significantly more capable than its previous Opus models, based on the controversial Mythos model. Anthropic had previously alerted the government in April about Mythos's potential to compromise advanced cybersecurity systems in sensitive sectors like finance.

While Anthropic assured it would not release the Mythos model publicly, it has provided access to a select group of companies through its Project Glasswing cybersecurity initiative. The company maintains that Fable 5 included safeguards and restrictions to prevent misuse, despite the government's national security concerns.

“

As we have said publicly, we believe the Government should have the ability to block unsafe developments as part of a transparent, fair, clear and fact-based process. This action does not adhere to those principles.