<\/p>\n
The White House has ordered a halt to federal use of Anthropic\u2019s AI tools after the company refused to relax safety measures desired by the Pentagon, highlighting tension over ethical boundaries and military applications of AI technology.<\/p>\n<\/div>\n
The White House has ordered all federal agencies to stop using Anthropic\u2019s artificial intelligence services, after the company refused Pentagon demands to remove safety restrictions on its Claude model, a move the administration says creates a supply-chain risk to national security. According to AP reporting, Defence Secretary Pete Hegseth and President Donald Trump framed the company\u2019s refusal as endangering troops and compromising the military\u2019s operational needs. [6]<\/a><\/sup>,[2]<\/a><\/sup><\/p>\n The dispute hardened after Pentagon negotiators sought contract language that would permit broader military use of Anthropic\u2019s systems, including applications the company has said would enable mass domestic surveillance or fully autonomous weapons. Anthropic\u2019s leadership declined to waive those limits on ethical grounds, prompting the Pentagon to designate the firm a security risk and to cancel a roughly $200m arrangement, Axios reported. [2]<\/a><\/sup>,[7]<\/a><\/sup><\/p>\n President Trump amplified the confrontation on social media, posting on Truth Social: “The Leftwing extremists at Anthropic have made a DISASTROUS MISTAKE by trying to STRONG-ARM the Department of War and forcing them to follow their Terms of Service instead of our Constitution,” and adding, “WE will determine our Country\u2019s future \u2013 NOT some out-of-control, Radical Left AI firm led by people who don\u2019t understand what the real world is like.” The administration used that rhetoric to justify its directive to halt federal use of Anthropic tools. [6]<\/a><\/sup>,[2]<\/a><\/sup><\/p>\n Anthropic has pushed back, announcing plans to sue the Pentagon and insisting it will not abandon the safety guardrails it says are essential to prevent misuse. The company argues the new contractual language, while presented as a compromise, would in practice undermine the stated protections by leaving them vulnerable to broad waiver or reinterpretation. Axios and AP both quoted Anthropic executives saying the firm cannot in conscience accede to demands that would permit its models to be repurposed for indiscriminate surveillance or lethal autonomous systems. [4]<\/a><\/sup>,[6]<\/a><\/sup><\/p>\n With Anthropic sidelined, the Pentagon has moved to secure alternatives. OpenAI said it reached an agreement to provide models to the Defence Department under explicit “red lines” that prohibit domestic mass surveillance and reaffirm human responsibility for use of force, a framework OpenAI\u2019s CEO described as consistent with the safety boundaries advanced by other firms. Industry sources told Axios the department is exploring OpenAI as a replacement partner. [3]<\/a><\/sup>,[2]<\/a><\/sup><\/p>\n The clash has reignited broader tensions over how commercial AI companies balance ethical constraints against national security demands. Critics of the administration\u2019s action warned it risks chilling investment and could force major cloud and chip partners to reconsider ties with Anthropic, while former officials described the blacklisting as heavy-handed. Supporters of the ban argue that military planners need adaptable tools without legal or contractual fetters that could impede operations. The episode underlines the thorny policy choices governments face as advanced generative models become central to both civilian and defence capabilities. [5]<\/a><\/sup>,[2]<\/a><\/sup><\/p>\n Inspired by headline at:<\/strong> [1]<\/a><\/sup><\/p>\nSource Reference Map<\/h3>\n