Trump orders federal agencies to stop using Anthropic AI

ABC News reports that Trump directed every federal agency to immediately cease all use of Anthropic technology.[2] Trump also said agencies using Anthropic products, including the Department of War, would have six months to phase them out.[2] Hegseth said the Department of War would continue receiving Anthropic services for up to six months while transitioning to another provider.[2] The administration’s statements pair an immediate cease-use direction with a stated phaseout period; the evidence does not specify how agencies should reconcile those timelines in practice.[2]

Anthropic said negotiations reached an impasse over its requests to bar mass domestic surveillance of Americans and fully autonomous weapons.[3] It says it supports lawful national-security uses of AI apart from those two exceptions, and that the exceptions had not affected any government mission to its knowledge.[3] Anthropic defines fully autonomous weapons as systems that select and engage targets without humans in the loop, and says current frontier models are unreliable for that use.[4][3] The Pentagon said it had no intention of using Anthropic AI for mass domestic surveillance or autonomous kinetic operations, but said the company’s safeguards could jeopardize military operations.[2]

Hegseth said he was moving to designate Anthropic a supply-chain risk, told defense contractors they could not use its AI, and declared the decision final.[2] Anthropic said it would challenge any such designation in court.[3] It says a formal designation could affect contractors’ Claude use only on Department of War contract work, while access through its API, claude.ai, or its products would be unaffected.[3] Reuters reported that the designation could bar tens of thousands of contractors from using Anthropic AI on Pentagon work.[1]

Anthropic said Claude is used across the Department of War and other national-security agencies for intelligence analysis, modeling and simulation, operational planning, and cyber operations.[4] It said it was the first frontier AI company to deploy models on classified U.S. government networks and had supported American warfighters since June 2024.[3] ABC News reported that OpenAI announced a deal to deploy its AI models on the Department of War’s classified network.[2] Reuters reported that Pentagon agreements with major AI labs had ceilings of $200 million each during the preceding year.[1] Senate Armed Services Committee leaders urged Hegseth and Anthropic CEO Dario Amodei to extend negotiations and work with Congress on a solution.[2]