Pentagon pressures Anthropic on AI guardrails

The dispute centers on two key restrictions Anthropic has placed on its AI model, Claude: a prohibition on its use for mass domestic surveillance and in fully autonomous weapons systems that can kill without human intervention. Anthropic CEO Dario Amodei has stated the company "cannot in good conscience" remove these safeguards, arguing the technology is not reliable enough for such applications. In response, the Pentagon has threatened to designate Anthropic a "supply chain risk," a label typically reserved for foreign adversaries, which could severely damage its commercial business. Officials have also suggested invoking the Defense Production Act to compel the company to comply with their demand for "any lawful use" of its technology. This standoff escalated after Anthropic's technology, deployed via a partnership with Palantir, was reportedly used in a U.S. military operation in Venezuela. Anthropic was the first of four major AI companies, including OpenAI, Google, and xAI, to be awarded a $200 million contract and be cleared for use with classified military networks. The Pentagon's aggressive stance is part of a broader push to accelerate AI adoption for military dominance, as outlined in Defense Secretary Pete Hegseth's "AI-first" warfighting strategy. Hegseth has publicly stated, "we will not employ AI models that won't allow you to fight wars." In contrast, other AI firms like Elon Musk's xAI have already agreed to the Pentagon's terms for classified use. The controversy has drawn reactions from across the tech industry, with employees at Google and OpenAI circulating a petition in support of Anthropic's position against the use of AI for mass surveillance and autonomous weapons. The Pentagon's Undersecretary for Research and Engineering, Emil Michael, has publicly criticized Amodei, accusing him of having a "God-complex" and putting national safety at risk.