Dario Amodei Challenges Department of Defense’s Supply-Chain Risk Designation
On Thursday, Dario Amodei announced that Anthropic is planning to contest the Department of Defense’s decision to categorize the AI company as a supply-chain risk in court. He has criticized this designation as “legally unsound.”
Following a prolonged dispute over the extent of military control over AI systems, the DOD officially classified Anthropic as a supply-chain risk. This designation can prohibit the company from collaborating with the Pentagon and its affiliates. While Amodei emphasized that Anthropic’s AI technology will not be used for mass surveillance of Americans or fully autonomous weapons, the Pentagon believed it should have unrestricted access for “all lawful purposes.”
Amodei clarified that the majority of Anthropic’s customers are unaffected by the supply-chain risk designation.
“For our customers, this designation specifically pertains to the utilization of Claude by customers as a direct component of contracts with the Department of Defense, not all interactions involving Claude by customers with such contracts,” he stated.
In preparation for the upcoming legal battle, Amodei highlighted that the Department’s letter labeling Anthropic as a supply-chain risk is narrowly focused.
“The purpose of this designation is to safeguard the government rather than punish a supplier; in fact, the law mandates the Secretary of Defense to employ the least restrictive means necessary to protect the supply chain,” Amodei explained. “Even for Department of Defense contractors, the supply-chain risk designation does not and cannot restrict the usage of Claude or business relationships with Anthropic if they are unrelated to their specific contracts with the Department of Defense.”
Amodei also mentioned that productive discussions had been ongoing between Anthropic and the DOD until an internal memo he sent to staff was leaked. In the memo, Amodei criticized OpenAI’s collaboration with the Department of Defense as “safety theater.”
OpenAI Steps In
Subsequently, OpenAI replaced Anthropic in working with the DOD, causing discontent among OpenAI employees.
Amodei expressed regret over the leak in his statement, asserting that the company did not intentionally disclose the memo or instruct anyone else to do so. He stated, “It is not in our best interest to escalate the situation.”
Amodei acknowledged that the memo, written hastily in response to a series of announcements, including Anthropic’s removal from federal systems, the supply-chain risk designation by Defense Secretary Pete Hegseth, and the Pentagon’s agreement with OpenAI, did not reflect his thoughtful views. He described it as an outdated assessment of a challenging day for the company.
He concluded by emphasizing Anthropic’s commitment to providing essential tools to American soldiers and national security professionals during significant combat operations. As Anthropic supports U.S. operations in Iran, Amodei assured that the company would continue supplying its models to the DOD at minimal cost for as long as necessary.
Legal Challenges Ahead
Anthropic may challenge the supply-chain risk designation in federal court, potentially in Washington. However, the legal framework surrounding the decision complicates the process by limiting companies’ ability to contest government procurement decisions and granting the Pentagon wide discretion on national security issues.
Dean Ball, a former White House AI adviser critical of Hegseth’s treatment of Anthropic, noted the difficulty in challenging government decisions related to national security. He remarked, “Courts are generally hesitant to question the government on national security matters… The bar to challenge such decisions is high but not insurmountable.”