Anthropic seeks en banc rehearing over Pentagon risk label

The D.C. Circuit majority said it had “no reason to doubt” Anthropic’s “noble intentions” and still upheld the Pentagon’s label of the AI lab as a “supply chain risk.” Anthropic has asked for a rehearing en banc, in front of all active judges of the court.

The majority read “otherwise manipulate” in FASCSA by its dictionary sense: to operate or control something skillfully. The fight began when Anthropic refused to let Claude be used for domestic mass surveillance or fully autonomous weapons.

A Lawfare analysis argues the law’s history, built around Kaspersky, Huawei and ZTE, targets covert compromise of government systems. Judge Karen LeCraft Henderson’s dissent described the conduct at issue as “a contractor’s honest and upfront enforcement” of use restrictions.