post-thumb

The White House imposed export controls on Anthropic after 24 hours

White House Imposes Export Controls on Anthropic AI Models After Security Concerns

The Trump administration imposed export controls on Anthropic’s latest artificial intelligence models, Fable 5 and Mythos 5, after senior officials raised concerns that the systems could pose national security risks, according to officials cited in the report.

The action followed a 24-hour period of discussions between the White House and Anthropic CEO Dario Amodei. Administration officials said they urged the company to voluntarily restrict access to Fable 5 after receiving information that its safeguards could be bypassed. Amazon, an investor in Anthropic, reportedly raised concerns with the White House after being asked for feedback. Amazon declined to discuss specifics, saying governments sometimes seek its counsel on security issues.

Senior officials, including Treasury Secretary Scott Bessent, White House Cyber Director Sean Cairncross, and Commerce Secretary Howard Lutnick, participated in calls with Amodei. According to administration officials, Amodei defended the model’s safeguards and argued that the reported issue did not amount to a broad “universal jailbreak.” Anthropic later said no testers had found such a universal bypass and described the government’s action as disproportionate.

The White House said export controls were used only after Anthropic declined to remove the model voluntarily. A person close to Anthropic disputed that account, saying the company was given a short deadline without detailed information about the alleged threat.

The export controls barred use of the models by foreign nationals, prompting Anthropic to disable them for all customers to ensure compliance. Officials said the decision was based on national security authority and could be lifted if the company addresses the issue.

The dispute highlights growing tension between AI companies and federal officials over how to regulate advanced models. Anthropic has publicly supported stronger AI oversight, while some administration figures have criticized the company on policy and political grounds. Officials said prior disagreements were separate from the export-control decision.

Share: