As the controversy around the military's use of AI heats up, Hegseth and the CEO of Anthropic are scheduled to meet

Pete Hegseth is set to meet with Dario Amodei, chief executive of Anthropic, as the company remains the only major AI firm not fully supplying its technology to the Pentagon’s new internal AI network.

Anthropic, maker of the chatbot Claude, declined to comment on the meeting. However, Amodei has publicly raised ethical concerns about unchecked government use of AI, warning about the dangers of fully autonomous armed drones and AI-assisted mass surveillance capable of tracking dissent. In a recent essay, he wrote that a powerful AI system analyzing billions of conversations could detect and suppress “pockets of disloyalty.”

The Pentagon last summer awarded contracts worth up to $200 million each to Anthropic, Google, OpenAI and xAI, founded by Elon Musk. Anthropic was the first to be approved for use on classified military networks, working alongside partners such as Palantir. The other companies are currently operating only in unclassified environments.

By early this year, Hegseth was publicly highlighting only xAI and Google. In a January speech at SpaceX in South Texas, he said he would shrug off AI models “that won’t allow you to fight wars,” emphasizing that Pentagon AI systems should operate without “ideological constraints that limit lawful military applications.” He added that the department’s AI “will not be woke.” Shortly after, Musk’s Grok chatbot was announced as joining the Pentagon’s GenAI.mil network, even as it faced scrutiny for generating highly sexualized deepfake images. OpenAI later confirmed it would provide a customized version of ChatGPT for unclassified military use.

Anthropic has long positioned itself as more safety-focused than its rivals, particularly after its founders left OpenAI in 2021 to form the startup. During the Biden administration, the company voluntarily subjected its AI systems to third-party scrutiny to mitigate national security risks. Amodei has warned that AI-related dangers are closer in 2026 than they were in 2023, though he argues they should be managed pragmatically rather than through alarmism.

The company has also clashed with the Trump administration on other policy fronts. It publicly criticized proposals to loosen export controls on AI chips to China, even while maintaining a partnership with Nvidia. Trump’s AI adviser, David Sacks, accused Anthropic last year of pursuing a “regulatory capture strategy based on fear-mongering,” after cofounder Jack Clark wrote about balancing technological optimism with “appropriate fear.”

Anthropic has sought to maintain bipartisan ties, hiring former Biden officials and recently adding Chris Liddell, a White House official during Trump’s first term, to its board.

The dispute echoes earlier tensions over military AI, including the controversy around Project Maven, a Pentagon drone surveillance initiative that prompted employee protests at Google. Today, experts say AI’s role in defense is already deeply embedded and unlikely to recede. While lower-risk applications include administrative and analytical work, battlefield uses — particularly those involving lethal force or nuclear systems — carry significantly higher stakes. Military officials, analysts note, have been grappling with those risks and mitigation strategies for nearly a decade.