WASHINGTON — The U.S. Commerce Department has restricted access to some of Anthropic’s most advanced artificial intelligence models, directing the company to prevent foreign nationals from using the systems and prompting a broader suspension of access, according to statements from Anthropic and U.S. officials. The measures, disclosed this month, affect the company’s recently released Fable 5 and Mythos 5 models and were imposed under export-control authorities citing national security concerns.
The directive was issued by Commerce Secretary Howard Lutnick in a letter to Anthropic Chief Executive Dario Amodei, according to reports citing the correspondence. The order requires the company to obtain government authorization before providing the models to foreign nationals and warns of potential civil and criminal penalties for non-compliance. U.S. officials said the restrictions were intended to address concerns that the models could be used for intelligence or military-related purposes by foreign actors.
Anthropic said it received an export-control directive requiring the suspension of access to Fable 5 and Mythos 5 for foreign nationals both inside and outside the United States. Because of the breadth of the order, the company said it was forced to disable access to the models for all users while it works to comply. “The U.S. government, citing national security authorities, has issued an export control directive,” the company said in a statement.
According to Anthropic, the government believes there may be a method of bypassing safeguards in Fable 5 that could assist in identifying software vulnerabilities. The company said details of the government’s concerns were limited and argued that the reported issue was narrow in scope and not unique to its systems. Officials have not publicly released technical evidence supporting the restriction.
The move represents one of the most expansive U.S. efforts to limit foreign access to frontier AI models developed by an American company. The restrictions apply not only to users abroad but also to foreign nationals located within the United States, including some employees of Anthropic, according to company statements and published reports.
Industry reaction has been mixed. Some technology executives and cybersecurity researchers have urged the administration to reconsider, arguing that restrictions on advanced AI systems could affect research and security work. Others have supported tighter controls on powerful AI technologies with potential national security implications.
As of Sunday, Anthropic’s Fable 5 and Mythos 5 models remained unavailable, while discussions between the company and U.S. officials were continuing, according to reports. Other Anthropic models were not affected by the directive.
