US orders Anthropic to halt foreign access to top AI models
Anthropic says it has been ordered to suspend foreign access to its Fable 5 and Mythos 5 AI models over US national security concerns
- The US government has ordered Anthropic to suspend foreign access to Fable 5 and Mythos 5.
- Anthropic said the order was issued under export control rules.
- The government did not give specific details of its national security concerns.
- Anthropic understands the concern may involve a possible jailbreak method affecting Fable 5.
- Other Anthropic models are not affected.
- The company says it believes there has been a misunderstanding and is working to restore access.
The US government has ordered Anthropic to suspend access for all foreign nationals to two of its most advanced artificial intelligence models, Fable 5 and Mythos 5, citing national security concerns.
Anthropic said on Friday that it had received an export control directive from the US government requiring the company to halt access to the two models. The order did not provide specific details about the national security concerns involved, according to the company.
The move marks a sharp escalation in Washington’s scrutiny of advanced AI systems, as governments increasingly treat frontier models not only as commercial technology but also as potential national security assets.
According to Anthropic, the company understands that the US government believes it has become aware of a method to bypass, or “jailbreak”, Fable 5. The company said the order had forced it to abruptly disable Fable 5 and Mythos 5 for all customers to ensure compliance, while access to its other models would not be affected.
Axios reported that the US government under President Donald Trump was moving to block foreign governments, companies and non-US individuals from accessing Anthropic’s most advanced models.
The report said US Commerce Secretary Howard Lutnick had sent a letter to Anthropic chief executive Dario Amodei stating that Mythos 5 and Fable 5 would be subject to export controls covering use outside the United States and foreign nationals inside the country.
Anthropic said it believed there had been a misunderstanding and that it was working to restore access to the models as soon as possible.
The restriction comes shortly after the Trump administration asked leading AI developers to voluntarily submit their most capable models for government cybersecurity testing before public release.
The development also adds to regulatory pressure on Anthropic at a sensitive moment for the company.
Reuters had reported earlier in June that a months-long dispute between Trump administration officials and Anthropic appeared to be easing in several parts of the US government as the company prepared for a possible initial public offering.
However, the sudden suspension order, together with the Pentagon’s earlier move in March to list the company as a supply chain risk, suggests Anthropic may continue to face long-term government scrutiny as Washington tightens oversight of advanced AI technology.
