US Orders Anthropic to Disable Advanced AI Models for Foreign Nationals, Forcing Global Shutdown

The US government has triggered an unprecedented shutdown of one of the world's most advanced artificial intelligence systems, issuing an emergency export control directive that forces Anthropic to pull its newly released Fable 5 and Mythos 5 models offline globally.[1][3]

The directive, delivered to the San Francisco-based AI company late Friday afternoon, explicitly ordered Anthropic to suspend access to the models for all foreign nationals, citing unspecified national security concerns.[2][4]

Because Anthropic's infrastructure cannot reliably verify the citizenship of every user in real time—a restriction that would also apply to the company's own foreign-born researchers and engineers—the company concluded that selective compliance was impossible.[4][6]

As a result, Anthropic abruptly disabled Fable 5 and Mythos 5 for its entire customer base, plunging millions of users and enterprise integrations into the dark just days after the models launched to widespread acclaim.[3][6]

The sweeping order marks a historic escalation in Washington's strategy to contain artificial intelligence proliferation. For years, US export controls have focused heavily on the physical supply chain, restricting the sale of advanced semiconductors to foreign adversaries.[2][5]

Now, the administration has crossed a new threshold, applying export controls directly to the software layer and API access of a commercial AI model, effectively demanding that a US tech company police the nationality of its global user base.[5][7]

The immediate catalyst for the shutdown appears to be a cybersecurity vulnerability. According to Anthropic, the government believes it has identified a method to "jailbreak" or bypass the safety guardrails of Fable 5, allowing the model to read specific codebases and identify software flaws.[3][6]

Anthropic has fiercely contested the government's rationale. In a sharply worded public statement, the company argued that the government has only provided "verbal evidence of a potential narrow, non-universal jailbreak."[2][6]

The company's engineers reviewed the alleged vulnerability and concluded that the exploit is minor, relatively simple, and relies on capabilities that are already widely available in rival models, including OpenAI's GPT-5.5.[4][6]

"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," Anthropic stated, warning that applying this standard across the industry would "essentially halt all new model deployments for all frontier model providers."[2][6]

The clash exposes a deepening rift between the US defense establishment and Anthropic. Earlier this year, the Pentagon moved to designate the company as a "supply chain risk" after Anthropic sought to draw red lines around the military use of its technology for domestic surveillance and autonomous weapons.[3][4]

Fable 5 and its unrestricted counterpart, Mythos 5, were designed to be Anthropic's crown jewels. Mythos 5, in particular, was touted as having the strongest cybersecurity capabilities of any model in the world, restricted to a vetted group of cyber defenders and critical infrastructure operators.[3][5]

The sudden recall has sent shockwaves through the global tech sector. Developers who had integrated Fable 5 into their applications over the past week are now receiving error messages, forcing them to fall back to older, less capable models like Claude Opus 4.8.[6][7]

The directive also raises profound questions about the future of the global internet. Former White House officials have noted that enforcing such an order long-term would require AI companies to implement strict identity verification and citizenship checks for every user, fundamentally altering how digital services are delivered.[2][5]

As Anthropic scrambles to negotiate with the Commerce Department to restore access, the incident serves as a stark reminder of the fragile architecture of centralized AI. A single government decree, issued on a Friday evening, was enough to instantly sever access to a frontier technology for the entire world.[1][4]