Anthropic CEO Rejects Pentagon’s Claude AI Ultimatum

Dario Amodei, the CEO of Anthropic, has publicly declined a recent ultimatum from the U.S. Department of Defense. ^[1][2] The Pentagon requested that Anthropic allow its Claude AI model to be utilized for mass surveillance and the development of fully autonomous weapons systems. ^[2][3] Amodei rejected these demands on February 27, 2026, citing significant ethical concerns and safety risks associated with such applications. ^[1][2] The decision highlights a growing tension between Silicon Valley AI developers and military requirements. ^[3] Anthropic maintains that its mission focuses on building safe and steerable AI systems. ^[1] By refusing the Pentagon’s terms, the company reaffirms its commitment to constitutional AI principles, even at the cost of high-value government contracts. ^[2][3]