Anthropic Won’t Lift AI Safeguards Amid Ongoing Pentagon Dispute: CEO

Anthropic CEO Dario Amodei said Thursday that the company will not remove protections from its Claude AI model, escalating a dispute with the U.S. Department of Defense over how the technology can be used in classified military systems.

The statement comes as the Department of Defense reviews its relationship with Anthropic and weighs possible consequences, including the cancellation of the company’s $200 million contract and possible invocation of the Defense Production Act.

“We cannot in good conscience accede to their request,” Amodei wrote, referring to the Pentagon’s January demand that AI contractors allow use of their systems for “any legal usage.”

While the Pentagon has since required AI vendors to adopt the standard “any lawful use” language in future agreements, Anthropic remained the only cross-border AI company to resist transferring control of its AI to the military.

On Wednesday, Axios First reported that the Pentagon had issued an ultimatum requiring unlimited military use of Claude. The deadline is reportedly Friday of this week.

“It is the prerogative of the ministry to select contractors that best fit their vision,” Amodei continued. “But given the substantial value Anthropic’s technology brings to our armed forces, we hope they will reconsider.”

In his statement, Amodei articulated the company’s position in line with U.S. national security objectives.

“I believe deeply in the existential importance of using AI to defend the United States and other democracies, and to defeat our autocratic adversaries,” he said.

He added that Claude is “widely deployed across the Department of War and other national security agencies for intelligence analysis, modeling and simulation, operational planning, cyber operations and more.”

War on AI

The dispute unfolds against broader concerns about how advanced AI systems behave in high-stakes military scenarios. In a recent study from King’s College London, OpenAI’s GPT-5.2, Anthropic’s Claude Sonnet 4, and Google’s Gemini 3 Flash deployed nuclear weapons in 95% of simulated geopolitical crises.

Speaking at SpaceX’s Starbase in Texas in January, Defense Secretary Pete Hegseth said the U.S. military plans to deploy the most advanced AI models.

That same month, reports it emerged that Claude was used earlier that month during a US operation to capture former Venezuelan President Nicolás Maduro. Amodei refuted the claim that Anthropic questioned specific military operations.

“Anthropic understands that the War Department, not private companies, makes military decisions,” he said. “We have never objected to specific military operations or attempted to limit the use of our technology in an ad hoc manner.”

Nevertheless, Amodei said the use of these systems for mass domestic surveillance or autonomous weapons is incompatible with democratic values ​​and poses serious risks.

“Today, cutting-edge AI systems are simply not reliable enough to power fully autonomous weapons,” he said. “We will not knowingly provide a product that endangers American combatants and civilians.”

He also addressed the Pentagon’s threat to label Anthropic as a “supply chain risk” while also potentially invoking the Defense Production Act.

“These last two threats are inherently contradictory: one labels us a security risk; the other labels Claude as essential to national security,” he said.

While Amodei has said the company will not comply with the Pentagon’s request, Anthropic has simultaneously revised its Responsible Scaling Policy and dropped a pledge to stop training advanced systems without guaranteed safeguards.

Robert Weissman, co-chairman of Public Citizen, said the Pentagon’s stance signals broader pressure on the tech industry.

“The Pentagon is publicly bullying Anthropic, and the public part is intentional, because they want to put pressure on this particular company and send a message to all the big tech companies and all the corporations that we intend to do and take what we want and not stand in our way,” Weissman told me. Declutter.

Weissman described Anthropic’s guardrails as “modest” and aimed at preventing “improper surveillance of the American people or to facilitate the development and deployment of killer robots, AI-enabled weapons that could launch lethal attacks without humans saying so.”

“Those are the most sensible and modest guardrails you can imagine when it comes to this powerful new technology.”

Regarding the Pentagon’s threat to designate Anthropic as a “supply chain risk,” Weissman called this a potentially crushing punishment from the government, and argued that this would put pressure on other AI companies to avoid imposing similar limits.

“Individuals may use Claude, but none of the AI ​​companies, especially Anthropic, have business models based on individual use; they are looking for enterprise use,” he said. “This is a potentially crushing punishment from the government.”

While the Pentagon has not yet said whether it plans to follow through on its threat to terminate the contract or invoke the Defense Production Act, Weissman said the Pentagon is telling AI companies that it expects unrestricted access to their technology once it is deployed in government systems.

“The message from the Pentagon is: ‘We are not going to tolerate this, and we expect to be able to use the technology as invented for whatever purpose we want,’” Weissman said.

The Department of Defense and Anthropic did not immediately respond Decode requests for comments.