Anthropic has been forced to pull access to its most advanced AI systems for anyone outside the United States following a direct order from federal agencies. The company confirmed Friday that it received an export control directive at 5:21pm GMT instructing it to suspend access to Fable 5 and Mythos 5 for all foreign nationals—including those currently working inside the US, even Anthropic employees themselves.

What Happened

The timing couldn't be worse. Fable 5 only dropped this week, built on top of the Mythos architecture that government agencies have been quietly using to hunt down security vulnerabilities lurking in critical systems—some undiscovered for decades. The order hit so fast that Anthropic had to cut off access with almost no warning, leaving international users and overseas partners scrambling.

Why This Matters

Here's where it gets interesting from a hacker standpoint. Mythos 5 isn't just good at finding bugs—it's scary good. The model can identify software vulnerabilities that have evaded human researchers for years, making it invaluable for hardening US infrastructure. But the same capability in malicious hands? That's straight out of a cyberwar playbook. No wonder agencies moved to lock it down.

Anthropic's Pushback

Anthropic isn't rolling over quietly. The company reviewed the triggering report and concluded it referenced a limited code review and error correction capability—something GPT-5.5 from rival OpenAI can do too. They're arguing that blocking software used by hundreds of millions of people over this is overkill, especially given how thoroughly Fable 5's safety measures have been tested.

The Bigger Picture

This comes just weeks after Anthropic publicly proposed a coordinated pause on advanced AI development, warning that the technology is advancing so fast we risk losing control. Whether you think that's genuine concern or corporate posturing, the US government clearly isn't interested in slowing down its own access to these tools—it's just making sure nobody else gets near them.

Key Takeaways

  • Mythos 5 and Fable 5 are now blocked for all foreign nationals, including those legally in the US
  • The trigger appears to be vulnerability detection capabilities that multiple AI providers possess
  • Anthropic received only partial information about specific security concerns
  • This follows Anthropic's own call earlier this month for industry-wide development pauses

The Bottom Line

This isn't really about national security—it's about maintaining a competitive moat. When your own model can do what GPT-5.5 does but you still get locked down, the message is clear: the US wants AI supremacy, not AI safety. The irony of Anthropic preaching caution while fighting for broader access to their own tech? That's just the game now.