Pentagon Reverses Course, Resumes Use of Anthropic's Claude AI Months After Blacklisting

The Pentagon has ceased using artificial intelligence tools from Anthropic, months after designating the company a "supply chain risk" earlier this year, according to multiple sources who spoke to the BBC. The move marks a reversal of the company's recent widespread use within the Department of Defense.

Earlier this year, the Pentagon labeled Anthropic a supply chain risk, though the specific reasons were not disclosed. Despite this, sources said Anthropic's Claude model was recently widely used across the department, indicating that the blacklisting had not been fully enforced until now. The BBC was told of the development but no official statement has been released by the Pentagon or Anthropic.

The decision raises questions about the Pentagon's procurement processes and the security of AI supply chains. It remains unclear what prompted the initial risk designation or the timing of the cessation. Neither the Pentagon nor Anthropic has commented publicly, and it is not known whether other AI vendors will face similar scrutiny. The development could have implications for Anthropic's business with government agencies and for the broader adoption of AI in defense.