U.S. Gives Anthropic 90 Minutes to Block Foreign Access to Its Latest AI Models

In an unusually abrupt move, the Biden administration gave Anthropic just 90 minutes to comply with an order blocking foreign access to its newest AI models, Mythos 5 and Fable 5, according to a report in the Financial Times citing a source close to the company. The White House had already restricted these models for foreign governments and individuals, both abroad and within U.S. borders, under new export controls.

Commerce Secretary Howard Lutnick informed Anthropic CEO Dario Amodei in a letter that the models would be subject to these controls, effective immediately. Notably, the Commerce Department's own agencies had approved Fable 5 for release just days earlier after routine checks. The sudden reversal followed intelligence from Amazon, whose researchers allegedly manipulated Fable 5 into revealing sensitive data that could enable serious cyberattacks. White House officials, after reviewing the findings, decided the fastest fix was to cut off access for foreign governments, companies, and individuals entirely.

Anthropic was given no prior warning or specific complaints before the 90-minute ultimatum. The incident underscores the growing tension between rapid AI deployment and national security concerns, as regulators scramble to keep pace with increasingly powerful models.

Source: RIA Novosti