Musk targets Dario as Pentagon grills CEO over Claude AI

Mumbai: US Defence Secretary Pete Hegseth has reportedly summoned Anthropic CEO Dario Amodei to the Pentagon for a “high-stakes” and “tense” meeting over the US military’s use of the company’s Claude AI model, as Elon Musk publicly accused the AI firm of stealing training data. According to a report by Axios, citing a senior defence official, the meeting was “not a friendly” one, as Anthropic has refused to remove restrictions on its technology despite pressure from the Defence Secretary.

Claude is currently the only AI system deployed within classified US defence networks, under a $200-million pilot contract signed last year. However, in a January 9 memo, Hegseth asked AI companies to renegotiate terms to lift restrictions on their models.

Anthropic has declined to fully remove safeguards, including limits on mass surveillance of Americans and on the development of fully autonomous weapons. Defence officials warned the company could be labelled a “supply chain risk,” potentially voiding contracts and preventing other Pentagon partners from using Claude, though an Anthropic spokesperson described the discussions as “productive.” Replacing the company would be complex due to Claude’s deep integration into defence systems, the report said.

Meanwhile, the Pentagon has signed agreements with xAI and is close to a deal with Google for its Gemini model, increasing pressure on Anthropic, The New York Times reported.

Ahead of the meeting, Anthropic alleged that three Chinese AI firms used chatbots to siphon millions of Claude outputs to train their own models. Musk responded on X, accusing Anthropic of stealing training data “at massive scale” and mockingly referring to the firm as “MisAnthropic.”