The controversy arrives as artificial intelligence increasingly becomes a geopolitical issue. Even as Washington restricts access to advanced AI systems over national security concerns, French President Emmanuel Macron has said that India and France are exploring deeper cooperation in AI, highlighting competing approaches to the future of global AI development and access.
Anthropic on Saturday announced that it has been forced to suspend access to its advanced AI models, Fable 5 and Mythos 5, after receiving an export control directive from the US government.
According to the company, the order requires all access to the two models to be blocked for foreign nationals, regardless of whether they are located inside or outside the United States. The restriction also applies to foreign-national employees working at Anthropic.
"The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance," the company said, adding that access to all other Anthropic models remains unaffected.
Breaking: As US orders Anthropic to disable AI models for foreign nationals, French President Macron says India, France looking at cooperative AI pic.twitter.com/TenMyOBTr4
— Sidhant Sibal (@sidhant) June 14, 2026
Anthropic said it received the directive at 5:21 pm ET and that the government did not provide detailed information about the alleged national security concern.
However, the company believes the action stems from concerns that a method has been discovered to bypass, or "jailbreak", the safety safeguards built into Fable 5.
A jailbreak is a technique used to circumvent an AI model's restrictions and obtain responses that would normally be blocked.
Anthropic said it reviewed the reported technique and found that it exposed only a small number of previously known and relatively minor vulnerabilities. The company further argued that similar capabilities can be found in other publicly available AI models without requiring any sophisticated bypass.
Defending the model's safety framework, Anthropic said Fable 5 underwent extensive testing before launch, including thousands of hours of red-team exercises conducted with the US government, the UK's AI Safety Institute (AISI), private organisations and internal teams.
According to the company, those tests showed that Fable's safeguards were substantially more effective than those of previously deployed models. Anthropic also said testers were unable to discover any "universal jailbreak" capable of broadly bypassing the model's protections.
The company acknowledged that perfect jailbreak resistance is unlikely to be achievable for any AI developer. Instead, it adopted what it calls a "defence-in-depth" strategy, designed to make jailbreaks either highly limited or extremely costly to develop while combining safeguards with continuous monitoring.
Anthropic said this strategy also explains its requirement for 30-day retention of customer data on Fable 5, a policy intended to help researchers identify and mitigate jailbreak attempts.
The company further claimed that the government has so far provided only verbal evidence of what appears to be a narrow, non-universal jailbreak.
According to Anthropic, the reported technique essentially involved asking the model to review a software codebase and identify vulnerabilities. The company argued that such capabilities are already widely available in other frontier AI models and are routinely used by cybersecurity professionals.
Anthropic also said it reviewed a report that it believes formed the basis of the government's directive and concluded that the demonstrated capabilities were broadly comparable to those available in other leading AI systems, including OpenAI's GPT-5.5.
While complying with the directive, Anthropic strongly disagreed with the decision.
"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," the company said.
Anthropic warned that if the same standard were applied across the AI industry, it could effectively halt the deployment of new frontier AI models.
The company added that while governments should have the authority to block genuinely unsafe AI systems, such decisions should be made through a transparent, fair and evidence-based process grounded in technical facts.
For now, Anthropic said it is complying with the legal order while working to restore access to Fable 5 and Mythos 5 as quickly as possible.
"We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible," the company said.
