Anthropic reportedly back in negotiations with the Pentagon

The CEO of Anthropic, Dario Amodei, is re-engaging in conversations with the US Department of Defense following a stall in discussions on Friday concerning the military's adoption of the company's artificial intelligence (AI) tools.

As reported by The Financial Times, Amodei is negotiating with Emil Michael, the under-secretary of defense for research and engineering, striving to finalise an agreement on how the Pentagon could utilise Anthropic's Claude models.

Talks broke down on Friday when President Donald Trump ordered federal agencies to cease using Anthropic's technologies. 

Additionally, Defense Secretary Pete Hegseth indicated he might classify the company as a supply-chain threat to national security.

Recently, Michael criticised Amodei, labelling him a "liar" with a "God complex," in a X post.

If a new contract is secured, it would enable the US military to maintain its usage of Anthropic's AI, which has reportedly been employed in conflicts involving Washington and Iran.

Claude was the first significant model integrated into the government’s classified networks, thanks to a $200 million contract granted by the DoD to Anthropic.

However, the company later insisted on assurances that its technologies would not be used for domestic spying or autonomous weaponry. 

The Pentagon insisted that the military retain the right to deploy the technology for any lawful use.

In a memo shared with FT on Friday, Amodei allegedly informed employees that towards the conclusion of negotiations, the Pentagon proposed agreeing to Anthropic's conditions if they omitted a particular phrase concerning "analysis of bulk acquired data" — a line he mentioned "perfectly represented the issue we were most worried about."

Amodei also included in his memo that communications from the Pentagon and OpenAI, which secured a fresh agreement with the Defense Department on Friday, were either "outright untruths about these matters or sought to create confusion around them."

The announcement of OpenAI's contract with the Pentagon, occurring shortly after the White House condemned Anthropic, sparked public criticism.