Pentagon stops using Anthropic AI tools after blacklisting company, BBC told

The Defense Department has completely stopped using Anthropic's artificial intelligence products, an unnamed official told the BBC. The disclosure follows Defense Secretary Pete Hegseth's February decision to label the company a national security supply-chain risk. He had set a six-month window for removing its tools, but the BBC says it is unclear why the process ran past that point.

The label came after Anthropic declined a Pentagon demand to strip safety limits from its Claude model and allow unrestricted military use. The company cited fears about mass surveillance and autonomous weapons, called the designation unprecedented and unlawful, and sued the Trump administration. Anthropic declined to comment on the latest statement.

Sources including former defense officials and contractors told the BBC that Claude remained in use until very recently for research, analysis and intelligence work, including in operations against Iran. It was built into Maven, a Palantir-run platform that organizes satellite imagery, drone footage and other data, and was used for tasks such as identifying possible targets. Lauren Kahn of Georgetown's Center for Security and Emerging Technology said such systems are not simple to swap out once embedded.

Summarized by PolitiKool from BBC’s reporting. Read the full story at BBC →