Pentagon Says It Has Stopped Using Anthropic's AI, BBC Reports
A U.S. Defense Department official told the BBC the Pentagon has stopped using Anthropic's products — but sources say Claude was still in use until recently. Here's what we know.
The U.S. Department of Defense says it no longer uses AI products from Anthropic, the company behind the Claude AI models. But according to the BBC, the full picture is more complicated.
What the Pentagon said
In a statement given to the BBC on Monday, October 5, a Defense Department official said the department has ceased the use of Anthropic products. Anthropic declined to comment.
How we got here
Anthropic's tools had been used by the U.S. government and military since 2024, and it was the first advanced AI company to have its tools deployed in government agencies doing classified work.
Earlier this year, according to reports, the Pentagon pressed Anthropic to remove safety restrictions from Claude and give the military unrestricted access. Reports say the company refused to drop its safeguards on autonomous weapons and mass domestic surveillance.
In February, Defense Secretary Pete Hegseth designated Anthropic a "national security supply-chain risk" and announced that the Pentagon would stop using its products by late August. It isn't clear why the change appears to have taken longer.
The court fight
Anthropic did not accept the designation. In March it sued the Defense Department in federal court in California and filed a second challenge in Washington, D.C. The two cases have gone in different directions:
| Date | Court | What happened |
|---|---|---|
| March 2026 | N.D. California | Judge Rita Lin granted a preliminary injunction while the case continued; the Pentagon's technology chief said the ban still stood under a different statute |
| August 27, 2026 | N.D. California | Judge Lin ruled the designation unlawful, citing First Amendment retaliation, a denial of due process and "arbitrary and capricious" action |
| September 25, 2026 | D.C. Circuit | A divided panel, 2-1, upheld the designation under a separate procurement law |
The two rulings concern different statutes, which is why they point in opposite directions. Anthropic has said it is considering all options, including further review. The Pentagon remains free to choose which AI vendors it uses, which is consistent with its statement that it no longer uses Claude.
The conflicting accounts
Multiple people familiar with the matter told the BBC that, as recently as the week before the statement, Claude was still being used by the department in research, analysis and intelligence gathering, as well as in military operations against Iran. According to reports of the story, the model was running inside Palantir's Maven Smart System, a data and targeting platform used by the U.S. military.
These accounts come from unnamed sources cited by the BBC, and the Pentagon's statement doesn't address them directly.
Why it matters
The dispute goes to the heart of a growing debate: who decides how AI is used in warfare — the governments that buy the technology, or the companies that build it? AI firms increasingly publish rules about what their models may and may not be used for, and this case shows how those rules can collide with military demands.
It also raises practical questions about how quickly a large organisation can replace AI tools that are built into systems it relies on.
What we don't know yet
- Exactly when Claude was removed from Pentagon systems.
- Which AI tools are replacing it.
- Whether the dispute affects Anthropic's work with other U.S. government agencies.
We'll update this story as more information becomes available. The court cases are the part most likely to move next.
The reality check
Much of what is known comes from unnamed sources in a single BBC report, and the Pentagon's short statement does not say when Claude was removed. The legal picture is clearer than the operational one: the court record is public, while how the military actually uses AI tools is largely not.
Should AI companies be able to limit how the military uses their technology? Share your view.
Sources: BBC report via Capital FM, MyJoyOnline, AI Weekly, TechCrunch on the August ruling, Nextgov on the D.C. Circuit ruling, Breaking Defense. Photo: U.S. Department of Defense (public domain).
