US Orders Anthropic to Suspend Advanced AI Models for All Foreign Nationals

The U.S. government has issued an unprecedented export control directive ordering AI developer Anthropic to immediately suspend access to its most advanced artificial intelligence models for all foreign nationals. The sudden mandate, which cites national security concerns, marks a historic escalation in Washington's efforts to restrict global access to frontier AI technology.[1][2]

The order, delivered by the Commerce Department late Friday afternoon, specifically targets Anthropic's newly released Claude Fable 5 and Mythos 5 models. In response, the company announced it would "abruptly disable" the models for all users worldwide to ensure compliance, as the directive applies to any non-U.S. citizen, whether located inside or outside the United States—including Anthropic's own foreign national employees.[1][4]

Amazon Web Services (AWS), which hosts Anthropic's models for enterprise clients, confirmed late Friday that it had revoked access to the Fable 5 and Mythos 5 models for all users across all regions at Anthropic's request. Access to the company's older models, such as Claude Opus 4.8, remains unaffected by the export control directive.[1][5]

The government's intervention stems from what it describes as a newly discovered method for bypassing, or "jailbreaking," the safety guardrails built into Fable 5. According to Anthropic, federal officials believe this bypass could allow the model to be used for identifying software vulnerabilities, potentially handing foreign adversaries a powerful automated tool for cyberattacks.[1][3]

Anthropic has strongly pushed back against the government's assessment, characterizing the situation as a "misunderstanding." The company stated that it reviewed a demonstration of the alleged jailbreak technique and found it only identified a small number of previously known, minor vulnerabilities. Furthermore, Anthropic argued that other publicly available AI models can discover these same vulnerabilities without requiring any bypass at all.[3][4]

"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 said in a statement, warning that if this standard were applied across the industry, it would "essentially halt all new model deployments for all frontier model providers."[1][4]

The timing of the directive is particularly disruptive, coming just days after Anthropic launched Fable 5 and Mythos 5 to widespread industry acclaim. Fable 5 was introduced as the first "Mythos-class" model cleared for general use, boasting state-of-the-art performance in software engineering and complex reasoning. To mitigate risks, Fable 5 was equipped with safety classifiers designed to silently route sensitive queries—such as those related to cybersecurity, biology, or chemistry—to the older, less capable Opus 4.8 model.[6]

Mythos 5, the unrestricted version of the same underlying model, was already heavily gated. Anthropic had limited its access exclusively to a vetted group of cyber defenders and critical infrastructure operators, acknowledging that the model possesses the "strongest cybersecurity capabilities of any model in the world."[6]

The Commerce Department's action represents a fundamental shift in U.S. tech policy. For years, Washington's export controls have focused heavily on the physical supply chain—specifically the advanced semiconductors and manufacturing tools required to train AI models. By directly restricting foreign access to the software itself, the government is crossing a new regulatory threshold.[1]

Industry analysts warn that enforcing such a directive could fundamentally alter the architecture of the internet. Dean Ball, a former White House official, noted that complying with the order would likely require AI providers to implement strict citizenship verification systems, effectively forcing users to prove their nationality before accessing advanced software.[5]

The abrupt suspension also arrives against a backdrop of existing friction between Anthropic and the U.S. government. Earlier this year, the Department of Defense labeled the company a "supply chain risk" after Anthropic refused to allow its AI models to be used by the military for domestic surveillance and fully autonomous weapons systems. Anthropic has since filed two lawsuits attempting to block that designation.[1]

As Anthropic works to negotiate with federal regulators and restore access, the broader tech industry is watching closely. The outcome of this standoff could set a definitive precedent for how the United States balances the commercial deployment of artificial intelligence against the imperatives of national security in an increasingly fractured digital landscape.