Lifetime Citizen Portal Access — AI Briefings, Alerts & Unlimited Follows
Lawmakers press Pentagon on Anthropic blacklist and AI procurement rules at Pentagon witness hearing
Loading...
Summary
Rep. Sara Jacobs and other members questioned Undersecretary Duffy about the Pentagon's designation of Anthropic as a supply-chain risk and whether less-restrictive options were tried; DOD defended the decision and cited vendor engagement and the need for trusted, flexible relationships for sensitive AI tools.
Rep. Jacobs pressed Undersecretary Michael Duffy on the Pentagon's February action designating Anthropic as a supply-chain risk after the company insisted on two contractual restrictions: no mass domestic surveillance and no use in fully autonomous weapons. Jacobs said the designation was troubling because Congress designed that authority to target adversarial foreign actors like Huawei, not U.S. firms, and she asked whether DOD explored narrower alternatives such as cancelling the contract, issuing DFARS guidance, or using waiver timelines.
Duffy said he had been consulted before the designation and defended the department's judgment. He argued that continually updated AI services introduce unique vulnerabilities and that DOD needs vendors willing to provide the operational flexibility required of sensitive defense tools: 'The nature of artificial intelligence introduces specific and very great vulnerabilities and the need for continued updates from the company' that introduce risk to acquisition, so we need trusted relationships,' he told the committee.
Jacobs contrasted the Anthropic designation with the department's continuing relationship with other AI vendors, noting OpenAI had secured a classified Pentagon deal with similar restrictions and that OpenAI publicly said Anthropic should not be designated a supply-chain risk. Jacobs called the move 'retaliation' if a government action treats two vendors differently for the same contractual terms.
Duffy replied that the department made its determination based on sustained engagement with the company and that, since the designation, alternative AI vendors had stepped forward to offer services without the same contractual limits. Several members warned the designation could discourage start-ups and frontier AI companies from working with the Defense Department.
The exchange did not produce a formal change in policy; members asked for further detail on whether DOD evaluated less-restrictive options and requested written explanations of the department's process. Duffy said the department would provide follow-up information.
The committee did not vote; the discussion was recorded as a formal oversight question and a follow-up request.
