Anthropic and the Pentagon are back at the negotiating table, FT reports

Dario Amodei, the CEO of Anthropic, has rekindled discussions with the U.S. Department of Defense following a recent impasse regarding the military's utilization of the company's artificial intelligence technologies, as reported by The Financial Times. Amodei is currently in dialogue with Emil Michael, the under-secretary of defense for research and engineering, in a final attempt to secure an agreement on the Pentagon's access to Anthropic's Claude AI models. The conversations collapsed last Friday when President Donald Trump ordered federal agencies to cease using Anthropic's tools. Furthermore, Defense Secretary Pete Hegseth expressed intentions to categorize the company as a supply-chain risk to national security. Tensions escalated last week when Michael publicly criticized Amodei, labeling him a "liar" with a "God complex" via a post on X. Reaching a new contract is crucial for the U.S. military to maintain its usage of Anthropic's technology, which has reportedly been integral to operations in Washington's conflict with Iran. Claude was the first significant AI model integrated into the government's classified networks through a $200 million contract with the DoD. However, Anthropic later sought assurances that its tools would not be employed for domestic surveillance or in autonomous weapon systems. The Pentagon, on the other hand, insisted that the military should be permitted to use the technology for any lawful purpose. In a memo reviewed by the Financial Times, Amodei informed his team that during the latter stages of negotiations, the Defense Department had indicated a willingness to accept Anthropic's terms if they removed a particular phrase regarding the "analysis of bulk acquired data," which Amodei expressed was a significant concern. Amodei also criticized the messaging from both the Pentagon and OpenAI—who announced a new agreement with the Defense Department on the same day the White House criticized Anthropic—as misleading. This timing spurred a backlash, resulting in a surge of downloads for Anthropic's app, while OpenAI's ChatGPT faced a notable increase in uninstalls. In response, OpenAI CEO Sam Altman acknowledged that his company "shouldn't have rushed" into the deal and promised to revise its safeguards concerning the Defense Department's use of its technology. Altman took to X to reiterate that he hopes the Department of Defense will extend similar terms to Anthropic as those agreed upon with OpenAI. Founded in 2021 by former OpenAI employees who departed over strategic disagreements, Anthropic positions itself as a "safety-first" alternative in the AI landscape. However, government officials have criticized the company for what they perceive as excessive caution regarding AI safety. A tech industry coalition, including members like Nvidia and Google, recently expressed concern over Hegseth's classification of a U.S. company as a supply-chain risk.