An official says the DOD has stopped using Anthropic's tools; sources: Claude was in use as recently as last week, including in military operations against Iran (BBC)

The US Department of Defence has discontinued the use of Anthropic's AI tools, according to an official who spoke to the BBC. The Pentagon ceased utilizing Anthropic's products, as stated in a departmental statement. Secretary of Defence Pete Hegseth had previously classified Anthropic as a supply chain risk on national security grounds in February, announcing the Pentagon's intent to discontinue its use by late August.

However, multiple sources familiar with the matter revealed that Claude, an Anthropic model, was still being employed by the department for research, analysis, intelligence gathering, and military operations against Iran just last week. Claude was particularly crucial in processing vast amounts of information, according to these sources, including former defence officials and contractors who had direct experience with the Pentagon's AI usage.

The reasons behind the Pentagon's decision to stop using Anthropic at this point remain unclear. Anthropic had been utilized by the US government and military since 2024, marking the first instance of an advanced AI company's tools being deployed in classified government agencies. Initially, Anthropic agreed to remove safety guardrails from Claude to grant the military unrestricted access to its AI tools, but when the company refused, citing concerns about mass surveillance and autonomous weapons, Hegseth designated Anthropic as a national security supply-chain risk, a status typically reserved for companies from nations posing a threat to the US.

Anthropic had filed a lawsuit against the Trump administration to overturn the designation, arguing it was unprecedented and unlawful. In the meantime, the Pentagon has entered into contracts with other major tech companies like Google, xAI, and OpenAI, with OpenAI tools gaining wider adoption in some military departments.