Anthropic to challenge DOD’s supply chain label in court

Darius Amodei he said on Thursday that Anthropic plans to challenge the Defense Department’s decision to name the AI ​​company a Supply chain risk in court, a name he has called “legally invalid.”

The statement comes hours after the department formally designated Anthropic a supply chain threat following a weeks-long debate over how the military should control AI systems. A supply chain risk designation could prevent a company from working with the Pentagon and its contractors. Amodei wrote firmly that Anthropic AI should not be used to monitor the American population or autonomous weapons, but the Pentagon believed that it should have access to all “legitimate” information.

In a statement, Amodei said most of Anthropic’s customers are not concerned about the mention of supply chain risk.

“In relation to our customers, it only applies to the use of Claude by customers as a direct part of contracts with the Department of Defense, not to use Claude with customers who have such contracts,” he said.

As a preview of what Anthropic could argue in court, Amodei said the department’s letter listing the company as a supply chain risk is limited.

“It exists to protect the government rather than to punish the seller; in fact, the law requires the Secretary of War to use it minimum mitigation measures required to achieve the goal of securing the supply chain,” Amodei said.

Amodei noted that Anthropic has been in good discussions with the department over the past few days, a discussion that some suspect has stalled as others suspect. memo inside He sent to the staff. In it, Amdodei referred to OpenAI’s activities with the Department of Defense as a “security forum”.

OpenAI has signed an agreement to work with the Department of Defense instead of Anthropic, a move that has sparked controversy among OpenAI employees.

Amodei apologized for the leak in a statement on Thursday, saying the company did not intentionally share or tell anyone else to do so. “It is not appropriate for us to escalate the problem,” he said.

Amodei said the memo was written within “a few hours” of several announcements, including President Social Truth’s announcement that Anthropic would be removed from the federal system, then Defense Secretary Hegseth’s security risk, and finally the Pentagon’s announcement of OpenAI. He apologized for his tone, calling it a “difficult day for the company” and said the memo did not reflect his “careful or thoughtful attitude.” It was written six days ago, he added, and is now a “temporary review.”

He concluded that Anthropic’s top priority is to ensure that America’s military and national security professionals maintain the necessary equipment in the midst of ongoing major conflicts. Anthropic is currently supporting some US projects in Iran, and Amodei said the company will continue to provide its samples to the Department of Defense at a “nominal cost” for “a long period of time to change.”

Anthropic could challenge their removal in federal court, possibly in Washington, but the law behind the decision makes it harder to compete because it restricts the way companies can challenge government decisions and make the Pentagon more aware of national security issues.

Or as Dean Ball — a former Trump-era White House adviser on AI who criticized Hegseth’s Anthropic approach — put it: “Courts are reluctant to say that the government is not a matter of national security…