Appeals Court Upholds Pentagon Blacklist of Anthropic in AI Guardrails Fight
A US appeals court allowed the Defense Department to keep Anthropic on a list that bars it from military contracts due to national security supply-chain concerns. The dispute centers on how much control AI companies should have over military uses of their models. Anthropic has objected to Claude being used for lethal autonomous warfare without human oversight or mass surveillance of Americans.
The appellate decision keeps Anthropic barred from defense contracts, after a lower court had found the Pentagon's designation unlawful. The conflict began when Defense Secretary Pete Hegseth met CEO Dario Amodei and sought looser model restrictions; Anthropic declined.
Anthropic's stated limits concern autonomous lethal force without human control and domestic mass surveillance. After the designation, the State Department shifted an internal chat tool to OpenAI, while Treasury and Health and Human Services started phasing out Claude. The Pentagon says contract terms, not retaliation, caused the rift.
The ruling could affect how much influence AI developers have over military applications of their systems. It may shape procurement choices for defense contractors and government agencies, potentially narrowing the tools available to them. It could also influence public expectations about human oversight and surveillance safeguards, though the legal fight remains unresolved.