Pentagon halts use of Anthropic AI tools after company labelled a supply-chain risk

The US Department of Defense has ceased using artificial-intelligence tools from Anthropic, an official told the BBC on Monday, months after the company was formally designated a national security supply-chain risk.

According to reporting by the BBC and MyJoyOnline, the decision follows a February designation in which US authorities labelled Anthropic a "supply chain risk." That designation was reportedly made after Anthropic declined to remove certain safety guardrails from its AI products.

Why the ban matters

The Pentagon’s move removes one of the private-sector providers whose models and services had been adopted for a range of defence-related work. The designation of Anthropic as a supply-chain risk signals heightened concern within US national security circles about the use of external AI systems and the control of their safety features.

Officials told the BBC the stop in use is a direct consequence of the earlier blacklisting. MyJoyOnline reported the same sequence, saying the Department of Defense is no longer using Anthropic’s tools after the company was put on the supply-chain risk list on national-security grounds.

What has been reported

• The BBC reported the Department of Defense had stopped using Anthropic’s AI tools and said an official provided that information.

• Both the BBC and MyJoyOnline said Anthropic was labelled a "supply chain risk" in February.

• The outlets say the label followed Anthropic’s refusal to remove safety guardrails from its products; the refusal is presented as the reason given for the designation.

What has not been reported

None of the reports include extended detail about which Anthropic products were in use by the Department of Defense, how broadly the tools had been deployed, the exact nature of the safety guardrails at issue, or the operational impact of the change. The outlets also do not quote Defence or Anthropic spokespeople directly in the material referenced here.

Context and implications

The move reflects growing tensions between government agencies seeking specific assurances or capabilities from AI vendors and private firms aiming to maintain safety features intended to limit misuse. Where governments have sought access, modification, or removal of constraints on models, companies have sometimes resisted on safety or ethical grounds; in this case, sources say that resistance led to a formal designation that cut off Pentagon use.

How this story may develop

The reports make clear the decision is recent and tied to the February designation; future reporting may add details about the operational scope of the ban, any appeals or legal responses from Anthropic, and whether other agencies follow similar steps. For now, the publicly reported facts are that the Department of Defense is no longer using Anthropic’s AI tools and that the company was officially identified as a supply-chain risk earlier this year.