Federal News
Pentagon Presses Anthropic on AI Military Use
March 17, 2026
The Pentagon is actively challenging AI company Anthropic over its refusal to permit full military use of its AI model Claude under a $200 million contract awarded in July 2025 for integration into classified defense networks. US Defense Secretary Pete Hegseth has issued a clear ultimatum demanding that Anthropic remove ethical restrictions that limit AI applications in autonomous weapons and surveillance, or risk contract termination. This standoff underscores a broader tension between national security imperatives and ethical constraints in defense AI procurement, with competitors like OpenAI and Google demonstrating greater flexibility to support all lawful military uses. Anthropic's recruitment of a weapons expert to prevent misuse of its AI highlights ongoing concerns about balancing innovation, security, and ethical responsibility in AI deployment.
- Why this matters: Procurement professionals should note the Pentagon's firm stance on unrestricted military use of AI technologies, signaling that vendors must align with defense operational requirements to maintain or secure contracts.
- The competitive landscape favors AI providers willing to support comprehensive defense applications, potentially impacting future contract awards and vendor selection.
- Organizations involved in AI integration for defense should evaluate ethical policy frameworks and risk mitigation strategies to meet evolving DoD expectations.
- This development may influence contract negotiations, compliance requirements, and partnership strategies within the defense AI ecosystem.
Partners must help 'warfighters win in any fight.'
— Sean Parnell, Pentagon spokesman
AI development needs 7balanced optimism and fear8.
— Jack Clark
He wants the company to approve full military use of its AI technology, including defense operations.
— Pete Hegseth
Agencies
Department of Defense, Pentagon, Defense Departmentโs Chief Digital and Artificial Intelligence Office, US Department of Defence, White House
Vendors
Anthropic, OpenAI, Google, xAI, Palantir
Contracts
$200 million, up to $200 million each
Locations
Sources
- AI showdown: Pentagon pushes $200M project as Anthropic resists 'killer AI' role · MSN · Mar 14
- US Defense Secretary Pete Hegseth warns Anthropic to allow full military use of its AI or risk losing Pentagon contract · MSN · Mar 16
- AI firm Anthropic seeks weapons expert to stop users from 'misuse' · BBC · Mar 17