Overview
- The D.C. Circuit on Wednesday denied Anthropic’s bid to pause its “supply‑chain risk” designation and set oral arguments for May 19.
- The label blocks the AI firm from Pentagon contracts and requires defense contractors to avoid using its Claude models in military work.
- A San Francisco judge on March 26 granted a separate injunction against a related order, prompting the administration to restore some access to Claude across the federal government, according to court filings.
- Government lawyers say Anthropic’s refusal to allow uses like mass domestic surveillance and fully autonomous lethal targeting creates operational risk, and the court said the balance of equities favors the military during an ongoing conflict.
- Anthropic argues the actions punish its speech and denied due process in a first-of-its-kind test of a procurement law used to cut off suppliers deemed national‑security risks, a fight that could push the Pentagon toward other AI vendors.