Foreign users can no longer access Anthropic’s newest AI models after the company received an export-control order from the Trump administration. Anthropic said the directive was based on national security concerns and applies to every foreign national, whether they are living in the United States or somewhere else. That also includes some of the company’s own employees.
The decision came only a few days after Anthropic launched Claude Fable 5 along with its Mythos 5 technology. The company described the release as its most capable public AI system so far. It also said the models included safeguards meant to prevent them from being used for harmful purposes, especially in cybersecurity.
In a statement, Anthropic said government officials told the company about what they believed was a possible way to bypass those safeguards. The company said it only received verbal information about what it called a limited jailbreak and argued that the issue did not justify pulling back a model already being used by a large number of customers. Anthropic added that it believes the situation resulted from a misunderstanding and that it hopes access can be restored.
The government has not publicly explained the full reason behind the order. Reports mentioned in the available information said officials were concerned that a group with links to China may have accessed the new model. Other reports said the government believed Fable 5 could be used to identify software vulnerabilities if its protections were bypassed. Those claims have not been independently confirmed.
This isn’t the first disagreement between Anthropic and the Trump administration. Earlier this year, the company refused to let the U.S. military use its AI models for domestic surveillance or fully autonomous weapons systems. After that, Anthropic was placed on a federal supply chain blacklist. The latest order adds another dispute between the company and the government.
Pentagon Chief Information Officer Kirsten Davies defended the move in a post on X, saying national security should come before business interests. Anthropic responded that it would follow the order even though it disagreed with the government’s assessment.
The restrictions could create problems for companies and researchers that depend on the newest Anthropic models. Foreign researchers working with the company may temporarily lose access, and employees in the United States who are working on visas could also be affected. Some businesses may need to change their workflow if parts of their staff cannot use the software anymore. Anthropic said the order only affects Fable 5 and Mythos 5, while its older AI models remain available.
The order also marks a change in how the United States is handling AI technology. Earlier export controls mainly focused on advanced computer chips. This time, the restrictions target the AI models themselves. Experts have warned that powerful systems with advanced cybersecurity abilities could be misused if they were used for cyberattacks, especially against industries that still depend on older computer networks.
This image is the property of The New Dispatch LLC and is not licenseable for external use without explicit written permission.
