Why did the U.S. government order Anthropic to restrict access to its advanced AI models?

The U.S. government ordered AI developer Anthropic to block foreign nationals from its Fable 5 and Mythos 5 models, citing unspecified national security concerns. This directive, which even included Anthropic's international employees, was prompted by alleged "minor findings" of potential AI safety "jailbreaks." To comply, Anthropic temporarily cut off access to both models for all customers, highlighting escalating tensions between innovation and governmental oversight in advanced AI.

How did Anthropic respond to the U.S. government's directive regarding its AI models?

Anthropic confirmed its compliance with the U.S. government's order to restrict access to its Fable 5 and Mythos 5 AI models. However, the company publicly expressed significant reservations, stating the government provided no specific details for its national security concerns. Anthropic denied any disclosure of a "concerning non-universal potential jailbreak" and highlighted its prior collaboration on safety protocols, raising questions about transparency and the basis of such directives.

What precedent does government intervention in AI development set for the industry?

This government action sets a significant precedent, demonstrating a willingness to directly intervene and dictate access to cutting-edge AI technologies based on unverified threats. It creates uncertainty for developers, potentially slowing innovation or pushing research to less regulated environments. The focus on "foreign nationals" also signals hardening geopolitical lines, likely accelerating sovereign AI initiatives globally and fragmenting the international AI landscape.

← Back to front page

Generative AI & ToolsSaturday, June 13, 2026

Anthropic cuts off Fable 5 and Mythos 5 access following government order

Original reporting by The Verge

In a sudden and sweeping directive, the U.S. government on Friday ordered AI developer Anthropic to immediately block all foreign nationals, both within and outside the United States, from accessing its advanced Fable 5 and Mythos 5 AI models. Citing unspecified national security concerns, the mandate was so broad it even included Anthropic’s own international employees. To comply, Anthropic took the drastic step of completely cutting off access to both models for *all* customers, a move that reverberates through the AI development community and highlights the escalating tension between innovation and governmental oversight.

Anthropic's pushback

While Anthropic confirmed its compliance with the unprecedented order, the company concurrently expressed significant reservations about its basis. In a public statement, Anthropic revealed that the government provided no specific details regarding its national security concerns, claiming any evidence of potential "jailbreaks"—methods to bypass AI safety guardrails—was presented verbally and described as "minor findings" that offered "no Mythos-specific uplift." The company highlighted its prior collaboration with US and UK governments on safety protocols and data retention policies designed to thwart malicious use, flatly denying any disclosure of a "concerning non-universal potential jailbreak that led to a harmful result." This standoff raises critical questions about transparency, the scope of national security powers in regulating cutting-edge AI, and the industry’s ability to navigate an increasingly scrutinized regulatory landscape.

Anthropic's unprecedented decision to fully restrict access to its Fable 5 and Mythos 5 models, prompted by a government directive citing unspecified national security concerns, marks a critical juncture for the AI industry. While complying, Anthropic's public statement reveals a stark disagreement over the severity of alleged vulnerabilities and the order's transparency. The company's lament regarding the lack of specific evidence, despite its own rigorous security protocols, highlights a rapidly escalating tension between innovation and state control in advanced AI.

Setting a Precedent

This unilateral government action establishes a significant precedent, demonstrating a willingness to directly intervene and dictate access to cutting-edge AI technologies based on unverified threats. The implications for trust, innovation, and international collaboration within the AI ecosystem are profound. Developers may now face increased uncertainty, potentially slowing research or pushing development towards less regulated environments. It also raises serious questions about due process and evidentiary standards governing future governmental oversight of dual-use AI.

Beyond domestic impact, the directive's explicit focus on blocking "foreign nationals" signals a hardening of geopolitical lines in the race for AI supremacy. This move will likely accelerate sovereign AI initiatives globally, as nations prioritize independent development and control over critical technologies, further fragmenting the global AI landscape. The incident underscores the urgent need for transparent, internationally coordinated frameworks to define and mitigate AI risks, lest ad hoc national security interventions become the norm, stifling progress and inadvertently fueling an opaque AI arms race.

Frequently asked questions

Why did the U.S. government order Anthropic to restrict access to its advanced AI models?: The U.S. government ordered AI developer Anthropic to block foreign nationals from its Fable 5 and Mythos 5 models, citing unspecified national security concerns. This directive, which even included Anthropic's international employees, was prompted by alleged "minor findings" of potential AI safety "jailbreaks." To comply, Anthropic temporarily cut off access to both models for all customers, highlighting escalating tensions between innovation and governmental oversight in advanced AI.
How did Anthropic respond to the U.S. government's directive regarding its AI models?: Anthropic confirmed its compliance with the U.S. government's order to restrict access to its Fable 5 and Mythos 5 AI models. However, the company publicly expressed significant reservations, stating the government provided no specific details for its national security concerns. Anthropic denied any disclosure of a "concerning non-universal potential jailbreak" and highlighted its prior collaboration on safety protocols, raising questions about transparency and the basis of such directives.
What precedent does government intervention in AI development set for the industry?: This government action sets a significant precedent, demonstrating a willingness to directly intervene and dictate access to cutting-edge AI technologies based on unverified threats. It creates uncertainty for developers, potentially slowing innovation or pushing research to less regulated environments. The focus on "foreign nationals" also signals hardening geopolitical lines, likely accelerating sovereign AI initiatives globally and fragmenting the international AI landscape.

Intro and outro generated by Printing Press AI from the source article above. Always consult the original reporting for verbatim quotes and primary sources.