Anthropic, Pentagon resume talks on ‘high stakes’ AI defense deal

Anthropic CEO Dario Amodei is reportedly back in talks with Pentagon for new AI defence partnership after the breakdown of negotiations with the US Defense Department which made the tech company a pariah.

As reported by Financial Times (FT), Amodei has been involved in renewed discussions with Emil Michael, under-secretary of defence for research and engineering, aiming to resolve the dispute revolving around Pentagon’s controversial access to Anthropic’s AI tools.

According to sources privy to the matter, if Anthropic succeeded in securing a new contract with Pentagon, it would allow the US military to deploy its AI models over classified networks and slash the company’s “supply chain risk” designation.

In late February, Amodei refused to accept the Pentagon’s AI military proposal on the grounds of controversial clauses, including lethal autonomous weapons and mass surveillance of American citizens.

As a result of Anthropic’s defiance over AI ethics, Pentagon designated the company “supply chain risk to national security” and criticised Amodei as a “liar” with a “God complex.”

Later, OpenAI struck a deal with the Defense Department and received heavy backlash over compromising AI safety guardrails.

“Near the end of the negotiation the [department] offered to accept our current terms if we deleted a specific phrase about ‘analysis of bulk acquired data’ which was the single line in the contract that exactly matched this scenario we were most worried about. We found that very suspicious,” wrote Amodei in an internal memo to staff on Friday, which was first reported by The Information and seen by the FT.

Both Anthropic and the Pentagon have declined to comment.