Anthropic has temporarily disabled access to its newest AI models, Fable 5 and Mythos 5, after receiving a directive from the U.S. Commerce Department that bars the company from providing the models to any foreign national.
According to Anthropic, the restriction applies not only to people outside the United States but also to foreign citizens living within the country, including some of the company’s own employees. Because of the broad scope of the order, the company said it had no practical way to limit access selectively and instead disabled the models for everyone. Its existing Claude models, including Claude Opus 4.8, remain available.
Anthropic Calls the Decision a Misunderstanding
The company stated that it received the directive without detailed explanations about the national security concerns behind the decision. However, officials reportedly informed Anthropic that the action was triggered by the discovery of a technique capable of bypassing safeguards protecting Mythos through the Fable 5 interface.
Anthropic argues that the reported jailbreak is highly limited and only exposes cybersecurity capabilities in a specific scenario rather than universally defeating the model’s protections. The company also claims similar techniques could potentially be used on other frontier AI models that are not facing comparable restrictions.
As a result, Anthropic believes the government’s response is disproportionate and could create an unrealistic standard for the entire AI industry.
Industry Experts React With Surprise
The decision has sparked widespread debate among AI researchers and policy experts. Some critics view the move as another example of increasing government pressure on Anthropic, pointing to previous disputes between the company and the U.S. government over defense contracts and military use of AI systems.
Others argue that Anthropic’s own messaging contributed to the situation. The company has repeatedly emphasized the powerful cybersecurity capabilities of Mythos and the extensive safeguards surrounding it. Some observers suggest that describing the technology as exceptionally sensitive may have encouraged regulators to treat it accordingly.
Potential Impact on Anthropic
The restriction could create uncertainty for Anthropic as it reportedly prepares for a public listing. Investors may question whether future government actions could limit the company’s ability to distribute advanced AI models globally.
Some analysts also warn that restricting access for foreign researchers working in the United States could discourage international talent from joining American AI companies, potentially benefiting competing AI ecosystems abroad.
What Happens Next
Anthropic says it is actively working with the U.S. government to resolve what it considers a misunderstanding and restore access to Fable 5 and Mythos 5 as soon as possible.
The incident also highlights a growing challenge for the AI industry: balancing national security concerns with the rapid development and global deployment of increasingly powerful artificial intelligence models. As governments introduce stricter oversight, similar disputes could become more common across the sector.
