Pentagon stops using Anthropic AI tools months after blacklisting company, BBC told
BBC News
Pentagon says it has stopped using Anthropic’s AI tools. That comes after months of using Claude inside its intel systems, despite a security blacklist.
Based on reporting by BBC News — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
The Pentagon says it has stopped using Anthropic’s AI tools, ending a messy split between public policy and what people inside the department say was still happening behind the scenes. A defence official told the BBC the department “has ceased the use of Anthropic products.”
That is a sharp turn from February, when Defence Secretary Pete Hegseth called Anthropic a supply chain risk and said the company would be phased out by late August. But people familiar with the matter told the BBC that Claude was still being used as recently as last week for research, analysis, intelligence gathering and even military operations against Iran.
Claude was not a side experiment. Sources said it sat inside Maven Smart System, the Pentagon’s main platform for organising intelligence and other data, which is run by Palantir. Analysts used Maven to pull together satellite imagery, drone footage and other visual material, then fed it into Claude and other models to help identify possible targets and prepare one-page briefs.
That integration helps explain why removal took so long. Lauren Kahn, a senior research analyst at Georgetown’s Center for Security and Emerging Technology and a former US defence official, said tools like this are not simple plug-ins. Once they are embedded, she said, pulling them out can be painful. The Pentagon also kept signing deals with other big AI firms, including Google, xAI and OpenAI, while Anthropic fought the designation in court.
The timing adds another layer. Hegseth set the six-month deadline on 27 February, one day before the US and Israel attacked Iran. Since then, Anthropic has called the Pentagon move “unprecedented and unlawful” and sued the Trump administration, while Dario Amodei has been back at the White House talking up AI safety with Donald Trump. That is a very Washington ending: blacklist the tool, keep using the tool, then call it a policy.
My take — AI-written commentary, not fact-checked reporting
This is what happens when a government falls in love with a tool before it has decided whether it trusts the company behind it. The real story isn’t just Anthropic’s fight with the Pentagon; it’s how quickly “national security risk” turns into “please don’t break Maven.” Washington loves grand security language right up until the spreadsheet gets angry.
Read more about this at: BBC News
