The U.S. Department of Defense has escalated its standoff with artificial intelligence developer Anthropic, giving the company a formal deadline to relax safety restrictions on its Claude AI model or risk losing critical military contracts and being labeled a “supply chain risk.”
According to reports, **U.S. Defense Secretary Pete Hegseth has instructed Anthropic CEO Dario Amodei to withdraw usage limitations on Claude by Friday, or face potential penalties that could include contract cancellation or invocation of the Defense Production Act — a rarely used federal authority that could compel the company to comply with government demands.
The core dispute centers on how artificial intelligence should be deployed in national defense. Anthropic has maintained firm restrictions on the use of its technology in certain military applications, specifically prohibiting mass domestic surveillance and fully autonomous weapons without human oversight — safety guardrails that reflect the company’s “safety-first” philosophical approach to AI development.
The Pentagon, however, argues that if Claude is to be used within classified or military systems, it must be made available for “all lawful purposes” without corporate-imposed limitations, a position that has heightened tensions over the proper balance between ethical AI governance and national security requirements.
Sources familiar with the negotiations say that while Claude remains in use on certain classified defense networks, the military is moving to ensure its unrestricted availability — a stance that places the Pentagon at odds with Anthropic’s ethical safeguards, potentially risking the company’s role as a defense supplier if no agreement is reached.
This dispute arises amid broader debates over how governments and private AI firms should collaborate as artificial intelligence becomes increasingly integrated into military operations, intelligence analysis, and strategic decision-making. Observers note that the outcome of this clash could set precedents for how ethical limitations on AI are treated in defense contracts going forward.
Sources
Reuters and Benzinga reporting – Anthropic and Pentagon at odds over AI usage policies and ethical guardrails.
Reuters / Axios report – Pentagon gives Anthropic until Friday to back down on AI safeguards.
San Francisco Chronicle – Pentagon threatens cancellation of $200M contract if Anthropic doesn’t ease safety restrictions.
Proactive Investors – Pentagon gives Anthropic deadline to drop AI restrictions or face removal from defense supply chain.
