WASHINGTON — Anthropic is facing new U.S. restrictions on foreign access to some of its most advanced artificial intelligence models after the government ordered the company to limit availability on national security grounds, prompting the firm to suspend access to the systems for all users while it complies with the directive.
The U.S. government issued an export-control order on June 12 requiring Anthropic to block access to its Fable 5 and Mythos 5 AI models for foreign nationals, including those located within the United States, according to statements from the company and reporting by Reuters. Anthropic said it received the directive without detailed public evidence supporting the national security concerns cited by officials.
“The U.S. government, citing national security authorities, has issued an export control directive,” Anthropic said in a statement, adding that the order required the suspension of access for foreign nationals and led the company to disable the models more broadly to ensure compliance.
According to Anthropic, the government believes a safeguard in Fable 5 could potentially be bypassed, allowing the model to assist in identifying software vulnerabilities. The company has disputed the severity of that risk, describing the reported issue as limited and arguing that comparable capabilities are available in other publicly accessible systems.
The restrictions mark one of the most significant U.S. interventions to date targeting access to specific AI models rather than hardware exports. Commerce Department officials have not publicly released detailed technical findings related to the concerns. Reuters reported that the administration’s directive applies to foreign governments, companies and individuals, while details of implementation remain limited.
The move has affected organizations that had been granted access to Anthropic’s advanced systems through restricted testing and research programs. Bloomberg News reported that some early participants retained access under specific arrangements, while other international organizations were informed that access would no longer be available. Anthropic has not publicly detailed all criteria governing those decisions.
The dispute has also led to discussions between Anthropic and U.S. officials over how future AI security risks should be assessed. According to Business Insider, representatives from the company and the White House have been working on a framework for evaluating vulnerabilities and determining when government intervention may be warranted.
As of Friday, access to Fable 5 and Mythos 5 remained suspended. Anthropic said other company models were unaffected, while discussions with U.S. authorities over compliance and security standards were continuing.
