AI Company Anthropic Refuses Pentagon Demand for Unrestricted Military Access to Claude

TLDR

• Anthropic and the U.S. Defense Department are locked in a dispute over a $200 million AI contract
• The AI company refuses to allow Claude technology for autonomous weapons and domestic surveillance operations
• Pentagon officials want unrestricted access to deploy AI regardless of company policies
• Defense Secretary Pete Hegseth stated the military won’t use AI models that restrict warfare capabilities
• The contract may be cancelled if the two sides cannot reach an agreement

Anthropic is in a heated dispute with the U.S. Defense Department over how the military can use its Claude AI system. The conflict threatens to end a contract worth up to $200 million.

The San Francisco-based AI startup has set strict limits on military use of its technology. Anthropic does not want Claude used for domestic surveillance or autonomous weapons targeting. The company requires human oversight for any weapons-related operations.

Pentagon officials reject these restrictions. They say the military should be able to use commercial AI technology however it wants. The only requirement should be following U.S. law, not company policies.

The contract was awarded last summer to integrate Claude into defense operations. Problems started almost right away. Anthropic’s usage terms ban domestic surveillance activities, which limits how agencies like ICE and the FBI can use the system.

Defense Secretary Pushes Back on AI Limits

Defense Secretary Pete Hegseth made the Pentagon’s position clear at a recent event. He said the military will not use AI models that restrict fighting capabilities. Sources confirmed he was talking about Anthropic.

The January 9 Defense Department AI strategy memo supports this approach. It states that commercial AI deployment should not be limited by company usage policies.

Anthropic CEO Dario Amodei has been vocal about AI safety concerns. He recently wrote that AI should support national defense except in ways that make America similar to autocratic countries. Amodei also criticized fatal shootings during immigration enforcement protests in Minneapolis.

Company Faces Pressure From Multiple Sides

The dispute puts Anthropic in a difficult position. The company is preparing for a public stock offering and recently entered talks to raise billions at a $350 billion valuation. Losing the Pentagon contract could hurt its national security business prospects.

Anthropic spent resources building relationships with defense and intelligence agencies. It was one of four major AI companies to win Pentagon contracts last year. Google, OpenAI, and Elon Musk’s xAI also received awards.

The Pentagon likely needs Anthropic’s help to use Claude effectively. The AI models are trained to avoid harmful actions. Only Anthropic staff can modify the system for military applications.

Amodei has clashed with White House AI czar David Sacks over regulation policy. Sacks accused Anthropic of being “AI doomers” trying to slow down competitors. The company denies this and says it maintains good relations with the administration.

An Anthropic spokesperson said Claude is used extensively for national security missions. The company stated it remains in productive discussions with the Department of War about continuing their work together.

Other AI companies working with the military have not faced similar public disputes. The outcome of this standoff could set precedent for how much control AI developers have over military use of their technology.

The post AI Company Anthropic Refuses Pentagon Demand for Unrestricted Military Access to Claude appeared first on Blockonomi.