A significant dispute has emerged between the U.S. Department of Defense (DoD) and AI company Anthropic over the latter's refusal to weaken safety protocols on its AI systems. The Pentagon's designation of Anthropic as a 'supply-chain risk' has jeopardized the company's government contracts, potentially impacting its projected 2026 revenue by billions.
The conflict centers on Anthropic's insistence on maintaining safeguards that prevent its AI from being used for autonomous weapons targeting and domestic surveillance. Legal experts suggest the Pentagon's actions may be legally questionable, citing internal inconsistencies and potential bias.
Key events include the Pentagon's pressure on AI companies to operate in classified settings without standard restrictions, culminating in a direct confrontation where Defense Secretary Pete Hegseth met with Anthropic CEO Dario Amodei. Following Anthropic's refusal to comply, the Pentagon issued an ultimatum, labeling the company a supply-chain risk. This led to several federal agencies, including State, Treasury, and Health and Human Services, ceasing their use of Anthropic's Claude AI. Defense contractors like Lockheed Martin have pledged to follow the DoD's direction.
Anthropic has responded by filing a lawsuit to block the Pentagon's designation, arguing it is unlawful and violates its rights. Microsoft has filed a brief supporting Anthropic, citing the potential for costly disruptions to its own products reliant on Anthropic's technology. The dispute highlights the growing tension between national security imperatives and the ethical deployment of advanced AI.