Anthropic Defies Pentagon Demands, Citing Ethical Risks in AI Defense Contracts

The confrontation between Anthropic and the Department of Defense (DoD) represents a fundamental clash between the "AI Safety" ethos of modern labs and the operational requirements of the U.S. military. Anthropic, founded on principles of "constitutional AI," is the lone holdout among major AI providers—including Google, OpenAI, and xAI—to resist the Pentagon's new internal network terms. CEO Dario Amodei’s statement that the company “cannot in good conscience accede” to the Pentagon’s demands underscores a growing rift between AI developers who prioritize ethical guardrails and a defense establishment seeking to maintain a technological edge.

At the heart of the dispute are two primary concerns: the potential for Claude, Anthropic’s flagship model, to be used for mass surveillance of American citizens and its integration into fully autonomous weapons systems. While the Pentagon, through spokesman Sean Parnell, has claimed it has no interest in illegal surveillance or human-out-of-the-loop weaponry, Anthropic maintains that the contract language provided by the DoD lacks the necessary legal and technical protections to prevent such outcomes. This resistance is particularly notable given that competitors like Google, OpenAI, and Elon Musk’s xAI have already agreed to supply their technology to the military’s new internal network, leaving Anthropic as the sole dissenter among the "Big Four" AI labs.

“

Anthropic, founded on principles of "constitutional AI," is the lone holdout among major AI providers—including Google, OpenAI, and xAI—to resist the Pentagon's new internal network terms.

The escalation of rhetoric from the Pentagon suggests a significant shift in how the U.S. government views the AI industry. During a high-stakes meeting between Defense Secretary Pete Hegseth and Amodei, military officials reportedly threatened to designate Anthropic as a "supply chain risk" or cancel its existing contracts. Most significantly, the DoD has floated the possibility of invoking the Defense Production Act (DPA). This Cold War-era law grants the President sweeping authority to prioritize government orders and direct industrial production for national defense. If invoked, it could theoretically force Anthropic to provide its models to the military regardless of the company’s internal ethical policies, setting a precedent that could fundamentally alter the SaaS and Cloud landscape for dual-use technologies.