Pentagon Clashes with Anthropic Over AI Weapons

- Anthropic CEO Dario Amodei, a prominent AI safety researcher, has built the company around a "Constitutional AI" approach, which embeds ethical guidelines directly into its systems to avoid harmful outputs. His public principles explicitly prohibit using their Claude AI for applications like autonomous weapons or mass surveillance. - The dispute stems from Defense Secretary Pete Hegseth's January 2026 "AI Acceleration Strategy," which mandates that all contracted AI models must be available for "all lawful purposes," directly conflicting with Anthropic's usage policies. Hegseth has publicly stated, "Department of War AI will not be woke. It will work for us." - The designation of a "supply chain risk" is a severe measure typically reserved for foreign adversaries like Huawei and could compel any company contracting with the Pentagon to sever ties with Anthropic. - While Anthropic has held firm on its ethical restrictions, other major AI labs with similar $200 million Pentagon contracts—including OpenAI, Google, and xAI—have reportedly agreed to the government's terms for "all lawful purposes" on unclassified networks. - The Department of Defense's official policy on autonomous weapons, Directive 3000.09, requires that systems be designed to "allow commanders and operators to exercise appropriate levels of human judgment over the use of force." However, it does not outright ban systems that can select and engage targets without direct human intervention. - Anthropic's Claude is reportedly the only frontier AI model currently operating on the Pentagon's classified networks, a deployment facilitated through a partnership with data analytics firm Palantir. - A high-stakes meeting between Anthropic CEO Dario Amodei and Defense Secretary Pete Hegseth is scheduled for this week to address the deadlock over the AI's use.