The US Department of Defence has stopped using Anthropic's AI tools, a Pentagon official confirmed to the BBC. The cessation comes months after Defence Secretary Pete Hegseth designated the company a national security supply chain risk in February and ordered a phase-out by late August.
However, multiple people familiar with Pentagon operations told the BBC that Anthropic's Claude model remained in active use as recently as last week for research, analysis, intelligence gathering and military operations against Iran. The tool was embedded in Maven Smart System, Palantir's primary intelligence platform used across the Defence Department to organize visual data including satellite imagery and drone footage. Analysts fed that data into Claude for tasks such as identifying potential military targets and compiling briefings for command.
Hegseth pressed Anthropic to remove safety guardrails from Claude and grant the military unfettered access. Anthropic refused, citing concerns about mass surveillance and autonomous weapons development. Hegseth then invoked a national security law typically reserved for companies based in countries deemed threats to the US.
Anthropic called the designation "unprecedented and unlawful" and sued the Trump administration to overturn it. The company had been used by US government and military agencies since 2024 and was the first advanced AI company with tools deployed in classified government work.
Lauren Kahn, a senior research analyst at Georgetown's Center for Security and Emerging Technology and former defence official, said the delay in removing Claude reflected integration challenges. "Once they become integrated it can be painful to remove them," Kahn said.
The Pentagon has since signed contracts with Google, xAI and OpenAI. A former senior defence official expressed concern about moving beyond Anthropic to OpenAI and other tools.
Relations between Anthropic and the White House appear to be thawing. Trump met with Anthropic chief Dario Amodei twice at the White House last week. After one meeting, Amodei said concerns about AI risks remained under discussion but expressed optimism about working with the president. Trump praised Amodei as "great" and "fantastic."