The US Division of Defence is not utilizing Anthropic’s AI instruments, an official has instructed the BBC, months after it designated the corporate a provide chain danger on nationwide safety grounds.
The Pentagon “has ceased using Anthropic merchandise”, the division official mentioned in a press release on Monday.
Secretary of Defence Pete Hegseth labelled Anthropic a provide chain danger in February, and introduced that the Pentagon would cease utilizing it by late August. It’s not clear why there was a delay.
A number of folks conversant in the matter instructed the BBC that as lately as final week Anthopic’s Claude mannequin was nonetheless being utilized by the division in analysis, evaluation and intelligence gathering in addition to in army operations in opposition to Iran.
It was seen as a essential software for sorting by way of huge portions of knowledge, the folks mentioned. They embody former US defence officers and contractors who’ve labored carefully with the Pentagon particularly on AI.
A spokesman for Anthropic declined to remark when requested in regards to the Pentagon’s assertion.
The small print of the division stopping all use of Anthropic’s instruments haven’t been beforehand reported.
The agency had been utilized by the US authorities and army since 2024 and was the primary superior AI firm to have its instruments deployed in authorities companies doing labeled work.
However earlier this 12 months, the Pentagon pressed it to take away its security guardrails from Claude and provides the army unfettered entry to its AI merchandise.
When Anthropic refused, citing issues of mass surveillance and autonomous weapons, Hegseth designated the corporate a “nationwide safety supply-chain danger” – a legislation sometimes reserved for firms primarily based in international locations that pose a risk to the US.
Anthropic said the move was “unprecedented and unlawful” and has sued the Trump administration in an effort to overturn the designation.
