Pentagon Issues Ultimatum to Anthropic Over AI Military Use
Defense Secretary Pete Hegseth has set a Friday deadline for Anthropic CEO Dario Amodei to open the company's artificial intelligence technology for unrestricted military applications, according to a source familiar with their meeting on Tuesday. Failure to comply could result in the termination of Anthropic's government contract, the source revealed, speaking on condition of anonymity due to lack of authorization to comment publicly.
Ethical Standoff and Military Demands
Anthropic, the creator of the chatbot Claude, remains the sole major AI firm not supplying its technology to a new U.S. military internal network. Amodei has consistently expressed ethical reservations about unchecked government AI use, highlighting risks such as fully autonomous armed drones and AI-assisted mass surveillance that could monitor dissent. In a recent essay, he warned, "A powerful AI looking across billions of conversations from millions of people could gauge public sentiment, detect pockets of disloyalty forming, and stamp them out before they grow."
The Pentagon has threatened additional measures beyond contract cancellation, including designating Anthropic as a supply chain risk or invoking the Defense Production Act to grant the military broader authority over its products, even without company approval. Despite the meeting being described as cordial, Amodei held firm on two non-negotiable boundaries: fully autonomous military targeting operations and domestic surveillance of U.S. citizens.
Broader Context and Industry Dynamics
This confrontation underscores ongoing debates about AI's role in national security, particularly concerning lethal force, sensitive information, and government surveillance. It aligns with Hegseth's pledge to eliminate what he terms a "woke culture" within the armed forces, emphasizing AI systems that operate "without ideological constraints that limit lawful military applications."
Last summer, the Pentagon awarded defense contracts worth up to $200 million each to four AI companies: Anthropic, Google, OpenAI, and Elon Musk's xAI. Anthropic was the first approved for classified military networks, collaborating with partners like Palantir, while the others currently operate only in unclassified environments. By early this year, Hegseth highlighted only xAI and Google, stating in a January speech at SpaceX that he was dismissing AI models "that won't allow you to fight wars." He later announced that Musk's chatbot Grok would join the Pentagon's GenAI.mil network, despite recent controversies over deepfake images.
Anthropic's Safety-Focused Stance and Political Tensions
Anthropic has long positioned itself as a more responsible and safety-minded AI company, founded in 2021 by ex-OpenAI employees. Owen Daniels, associate director at Georgetown University's Center for Security and Emerging Technology, noted, "Anthropic's peers, including Meta, Google and xAI, have been willing to comply with the department's policy on using models for all lawful applications. So the company's bargaining power here is limited, and it risks losing influence in the department's push to adopt AI."
During the AI boom following ChatGPT's release, Anthropic aligned with President Joe Biden's administration, volunteering for third-party scrutiny to mitigate national security risks. Amodei has cautioned about AI's potential catastrophic dangers but rejects the "doomer" label, advocating for "realistic, pragmatic" risk management. However, the company has faced tensions with the Trump administration, criticizing proposals to loosen export controls for AI chips to China and engaging in lobbying battles over state-level AI regulation. David Sacks, Trump's top AI adviser, accused Anthropic of "running a sophisticated regulatory capture strategy based on fear-mongering."
To navigate political divides, Anthropic has hired former Biden officials and recently added Chris Liddell, a Trump-era White House official, to its board. Amos Toh, senior counsel at the Brennan Center's Liberty and National Security Program, emphasized the need for congressional oversight, stating, "The law is not keeping up with how quickly the technology is evolving. But that doesn't mean DoD has a blank check."
The Pentagon has not commented on the development, initially reported by Axios, though a defense official confirmed the meeting between Hegseth and Amodei.
