US Government Orders Anthropic to Suspend Global Access to Its Most Advanced AI Models

In an unprecedented move, the Trump administration has issued an export control directive ordering Anthropic to immediately suspend all access to its most powerful AI models — Fable 5 and Mythos 5 — for foreign nationals, whether inside or outside the United States. To ensure compliance, Anthropic was forced to disable access for all of its customers worldwide.

What Happened

Anthropic released Claude Fable 5 and Mythos 5 less than a week before the government's order. The US Commerce Department, citing national security authorities, issued the export control directive on June 13.

According to reporting by Semafor, the trigger was a warning to the government that Fable 5 could be jailbroken — with a China-linked group reportedly having accessed the model. David Sacks, a Trump administration AI adviser, stated publicly on X that Anthropic was notified and declined to fix the issue before the export controls were imposed.

Anthropic disputes the severity of the threat: "We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people," the company stated. It called the situation a "misunderstanding" and said it is working to restore access as soon as possible. Access to all other Anthropic models remains unaffected.

Why AI Experts Are Concerned

Experts cited by Reuters noted that the Mythos model architecture, if misused, could dramatically accelerate sophisticated cyberattacks — particularly in sectors like banking that rely on legacy interconnected systems.

The restriction also adds complexity to the relationship between Anthropic and the Trump administration. Anthropic is currently suing the administration separately over being placed on a supply chain blacklist after refusing to allow US military use of its AI for domestic surveillance and fully autonomous weapons systems.

Broader Implications

The order affects:

• Foreign H1-B workers at US companies using Anthropic's tools
• Research institutions that collaborate with Anthropic globally
• Enterprises that have integrated Fable 5 or Mythos 5 into workflows
• Foreign-born Anthropic employees, whose access status remains unclear

The Pentagon's chief information officer publicly backed the decision, framing it as a matter of national priority over commercial concerns.

The move is the latest extension of US export control policy — previously applied to chips (Nvidia, AMD) — into AI software, signaling that advanced model access is now itself considered a controlled technology.

Read the full article on Al Jazeera