AI safety and research company Anthropic has told the Pentagon it will not agree to their demands to drop critical safety precautions and grant the U.S. military full access to their AI capabilities.
On Tuesday, Pete Hegseth’s Department of Defense reportedly threatened to cancel their $200 million contract with the San Francisco-based company and deem them a “supply chain risk” if they did not comply with the demands by Friday evening, according to Axios.
In a statement released on Thursday, Anthropic CEO Dario Amodei said his company “cannot in good conscience” comply with Hegseth’s requests despite the financial threats and hoped the secretary of defense would “reconsider” his position.
Anthropic’s AI program, Claude, is used for the military’s most sensitive work. The feud between Hegseth and Amodei is understood to have escalated over reports that Claude was used by the Pentagon in the Trump administration’s lightning invasion of Venezuela earlier in January.
The statement directly addressed Hegseth’s demands, noting the threat to designate them a “supply chain risk” is a label usually reserved for adversaries of the U.S. and has “never before” been applied to an American company.
Amodei also said Hegseth threatened to invoke the Defense Production Act to force the removal of the safeguards to allow military access. He said the two threats were “inherently contradictory” as “one labels us a security risk; the other labels Claude as essential to national security.”
Amodei’s statement said the company does not want AI to be implemented for mass surveillance of Americans or for the technology to be used to develop weapons that can be fired without human involvement, as it is “simply outside the bounds of what today’s technology can safely and reliably do”.
“Our strong preference is to continue to serve the Department and our warfighters—with our two requested safeguards in place,” Amodei said.
Those two safeguards state that while they support using AI for “lawful” foreign intelligence, deploying the technology for “mass domestic surveillance” of Americans is ”incompatible with democratic values.”
Amodei‘s statement points out that under the current law, the government can purchase “detailed records” of someone’s web browsing history and movements without obtaining a warrant, despite privacy concerns.
“Powerful AI makes it possible to assemble this scattered, individually innocuous data into a comprehensive picture of any person’s life,” Amodei said, noting it would happen “automatically and at massive scale.”
The CEO also said AI systems are currently “not reliable enough” to power fully autonomous weapons and they “will not knowingly” provide a product that could harm American military members and civilians.
Anthropic said it has offered to “work directly” with the Department of Defense on research and development to “improve the reliability of these systems, but they have not accepted this offer.”
Emil Michael, the Under Secretary of Defense for Research and Engineering, personally lashed out at Amodei on X on Thursday after the statement was released.
“It’s a shame that @DarioAmodei is a liar and has a God-complex,” Michael wrote. “He wants nothing more than to try to personally control the US Military and is ok putting our nation’s safety at risk.”
The post claimed the Department of Defense will “ALWAYS adhere to the law” but refused to “bend to whims of any one for-profit tech company.”
Hegseth reposted Michael’s comment, but has yet to make his own on his X account.
Michael kept posting, saying Anthropic had “dropped their own safety policy but thinks they should dictate ours” using the hashtag #GodComplex.
Democrat senator Elissa Slotkin posting on X “Why wouldn’t Secretary Hegseth say ‘No problem, we’re never gonna use AI at the Pentagon to conduct mass surveillance of American citizens?’”
Michael replied, claiming that Amodei “wants to override Congress and make his own rules to defy democratically decided laws. He is trying to re-write your laws by contract. Call @DarioAmodei to testify UNDER OATH!”
The Daily Beast has contacted the Department of Defense and Anthropic for comment.
The post AI Giant Tells Pentagon Pete Hegseth It Won’t Bow to His Demands appeared first on The Daily Beast.
