U.S. flags Anthropic as a security risk
The U.S. government labeled Anthropic’s Claude AI an “unacceptable risk” to military supply chains, a move that’s reignited Silicon Valley debate over defense ties and AI governance — Microsoft is publicly backing Anthropic in the dispute. (storyboard18.com)
The Department of Defense argued in a March 17 court filing that Anthropic’s safety limits and usage restrictions made the company an “unacceptable” wartime partner, explicitly questioning whether Anthropic could be a “trusted partner” in wartime operations. (nytimes.com) Defense Secretary Pete Hegseth formally designated Anthropic a supply-chain risk in early March, a decision the Pentagon said followed failed negotiations over the military’s ability to use Claude for “all lawful purposes.” (politico.com) President Joe Biden’s administration — via an order issued February 27 that instructed federal agencies to cease using Anthropic products — and the Pentagon’s designation have prompted Anthropic to file suit seeking to reverse the designation. (mayerbrown.com) Microsoft filed in court asking for a temporary restraining order to block the Pentagon’s ban, arguing a pause would “enable a more orderly transition” for military contracts; a coalition of retired military leaders also filed affidavits backing Anthropic. (cnbc.com) The Pentagon’s filings say Anthropic’s “red lines” — including limits on certain wartime uses and clauses that could let Anthropic disable capabilities — are the basis for the supply-chain designation. (forbes.com) Defense and contractor teams warn the Claude footprint is embedded in workflows and integrations, with some estimates saying a full replacement could take 12–18 months, even as DoD officials have told Congress they are confident replacements can be phased in within six months. (capacityglobal.com)
