DistantNews
Support us
🇧🇪 Belgium /Technology

U.S. Forces Anthropic to Disable Top AI Models: 'Highly Exceptional Government Intervention'

From VRT NWS · () Dutch

Translated from Dutch, summarized and contextualized by DistantNews.

At a glance

News Named sources New plan
  • The U.S. government has ordered AI company Anthropic to disable its two most powerful AI models, Mythos 5 and Claude Fable 5, for foreign users due to national security concerns.
  • Because the company cannot distinguish between foreign and domestic users in real-time, it has disabled the models for all customers, including its own employees.
  • Anthropic stated it was a misunderstanding and aims to restore access as soon as possible, while the U.S. Department of Commerce cited national security in its order.

The U.S. government has imposed an "exceptionally unusual" order on AI firm Anthropic, compelling it to disable its two most powerful AI models, Mythos 5 and Claude Fable 5, for foreign users. The directive, issued by the U.S. Department of Commerce on Friday evening, cites national security concerns.

Anthropic announced that because it cannot technically differentiate between foreign and domestic users in real-time, it has been forced to disable the models for all customers. This includes U.S. citizens and Anthropic's own employees. The company described the situation as a misunderstanding and expressed its intention to restore access as quickly as possible.

This marks the first time the U.S. government has mandated a domestic tech company to take such action regarding software usage. AI expert Tom Van de Weghe of VRT NWS described the situation as "highly exceptional," noting that while export controls on chips are common, government dictation of software access, even for a company's own foreign employees, is unprecedented. The inability to make real-time distinctions between user nationalities led to the system being locked for everyone.

This is highly exceptional. Export controls on chips towards countries like China are known, but that a government dictates who can use software, including the company's own employees with, for example, a Canadian or British passport, there is no precedent for that as far as I know.

— Tom Van de WegheDescribing the U.S. government's order to Anthropic as unprecedented.

Access to Anthropic's other AI models remains unchanged. Risky queries will be rerouted to an older model, Claude Opus 4.8, which is not subject to the order. Concerns about Mythos 5 arose earlier due to its ability to rapidly identify vulnerabilities in software security, raising fears it could become a powerful tool for cybercriminals. Anthropic had previously shared the system with select organizations managing critical infrastructure to allow them to patch weaknesses before exploitation.

Anthropic maintained that sufficient security safeguards were in place but acknowledged the possibility of "jailbreaks", circumventing software restrictions. The company stated that with perfect protection against jailbreaks currently seeming impossible, Fable 5 was designed with a layered defense strategy. However, the U.S. government's intervention underscores the growing concerns about the potential misuse of advanced AI technologies.

Because such a distinction cannot be made technically in real-time, the system then just locks up for everyone.

— Tom Van de WegheExplaining the technical reason for Anthropic disabling its AI models for all users.
DistantNews Editorial

Originally published by VRT NWS in Dutch. Translated, summarized, and contextualized by our editorial team with added local perspective. Read our editorial standards.

More Stories

From Our Blog