Enterprise technology leaders who rely on cutting-edge AI models face an unexpected disruption after Anthropic pulled access to its Mythos 5 and Fable 5 models under a direct order from the US government. The move, which the company says was triggered by a national security directive, has left users in what one described as an "absolutely nightmare scenario" — a situation that underscores the growing tension between AI deployment and government oversight.
The Security Directive
The directive came from the White House at 5.21pm Eastern time, ordering Anthropic to block access to its next-generation AI models to "any foreign national, whether inside or outside the United States, including foreign national Anthropic employees," according to Anthropic's blog post. Because the company currently has no way to apply that restriction selectively, it has universally pulled access to both models. Access to other Anthropic models is not affected.
Anthropic stated it is working to "restore access as soon as possible" and described the situation as a "misunderstanding." The company noted that it had previously worked closely with authorities in the US and the UK, and with multiple third-party agencies and internal teams, to ensure Mythos 5 and Fable 5 were safe to deploy. Fable 5's safeguards were described as "substantially more effective than those of any previously deployed model" by Anthropic.
The Jailbreak Concern
While few details have been released, Anthropic indicated that the US government is worried about a "potential jailbreak" — a method to bypass the model's safety guardrails. However, Anthropic characterized the vulnerability as "narrow," "non-universal," and not reason enough to pull access. In the company's words: "If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers."
| Model | Type | Access Status |
|---|---|---|
| Mythos 5 | Full, trusted-partner version | Pulled (universally) |
| Fable 5 | More restricted, public version | Pulled (universally) |
Fable 5 had only been released for a matter of days before access was revoked. It promised a significant step up in coding, reasoning, and agentic actions, and had undergone a thorough vetting process.
User Reaction
Users who had begun testing the models expressed dismay. According to one poster, the situation is an "absolute nightmare scenario." Another wrote, "this has to be a joke." There is speculation among the user community about what this implies for other AI models, including those from China, though no details are available.
What This Means for Enterprise AI Deployments
For enterprises evaluating or already using Anthropic's models, this incident highlights the risk of depending on frontier AI that can be withdrawn at a government's request. While Anthropic is working to restore access, the disruption is immediate and without a clear timeline. The company's disagreement with the government's assessment suggests ongoing uncertainty about the severity of the jailbreak and whether other models could face similar restrictions.
Organizations that had integrated Mythos 5 or Fable 5 into their workflows must now pause adoption and consider fallback options. The broader lesson is that even thoroughly safety-tested models can be pulled due to national security concerns, potentially affecting supply chain planning, logistics automation, and other enterprise processes that rely on advanced AI capabilities.
Sources:
