(Bloomberg) — Defense Secretary Pete Hegseth said the US doesn’t let artificial intelligence make lethal targeting decisions even as he called Anthropic PBC Chief Executive Officer Dario Amodei an “ideological lunatic” after a recent spat over military use of the company’s chatbot.
Hegseth again slammed Anthropic over its insistence that its Claude chatbot not be used for mass surveillance against Americans or in fully autonomous weapons operations. Those demands led the Pentagon in March to take the unprecedented steps of designating Anthropic a national-security threat and moving to bar its products from the military.
“On Anthropic they wouldn’t agree to our terms of service,” Hegseth said Thursday at a Senate Armed Services Committee hearing. “That would be like Boeing giving us airplanes and telling us who we can shoot at.”
Without referring to Amodei by name, he said Anthropic is run by “an ideological lunatic who shouldn’t have a sole decision-making over what we do.” Pressed by Senator Jacky Rosen, a Nevada Democrat, he said keeping a human in the loop in lethal targeting decisions is part of US terms of service. He didn’t respond directly when she asked if “there will always be a human in the loop.”
“We follow the law and humans make decisions,” Hegseth said. “AI is not making lethal decisions.”
Until recently, Anthropic was the only AI company approved for classified work. Its Claude Gov AI tools are used by Palantir Technologies Inc. in the Maven Smart System, an AI-enabled mission control platform that US armed forces have used in operations against Iran.
The clash with Anthropic has fanned fears that the US military will eventually allow artificial intelligence to pick and strike targets without any human oversight.
“The suggestion was made that somehow AI might be used without a human in the loop, which is a classic Anthropic talking point,” Hegseth said earlier in his testimony. “There is a human in the loop on decisions that are made, and the suggestion otherwise is to suggest that somehow AI is running targeting.”
Anthropic’s technology has drawn increased attention following the release earlier this month of its breakthrough Mythos model, which the firm has warned could pose a global cybersecurity risk and has limited access to a handful of Wall Street banks and tech companies for testing.
Trump administration officials have sought to make the system available to federal agencies, and Amodei met on April 17 with White House Chief of Staff Susie Wiles to discuss Mythos, among other topics.
More stories like this are available on bloomberg.com
#Hegseth #Calls #Amodei #Lunatic #Defends #Pentagon
