The White House and Pentagon are taking opposing positions on whether federal agencies should use Anthropic’s artificial intelligence systems, escalating a dispute over national security risks and government access to advanced AI tools.<\/p>\n
Pentagon Moves To Restrict Anthropic<\/p>\n
The Pentagon blacklisted Anthropic from military work after concerns about potential misuse of its AI for surveillance and autonomous weapons systems, ordering agencies to remove its products within six months.<\/p>\n
Defense Secretary Pete Hegseth criticized the company’s leadership, calling CEO Dario Amodei an “ideological lunatic” during a Senate hearing last week, and comparing the situation to “Boeing giving us airplanes and telling us who we can shoot at.”<\/p>\n
Pentagon officials have labeled Anthropic a “supply chain risk,” a designation typically reserved for foreign adversaries, and have ordered its removal from defense systems within months, with limited exceptions.<\/p>\n
President Donald Trump has signaled a more flexible stance, told<\/a> CNBC that a deal with Anthropic is “possible,” and praised its technical talent. <\/p>\n “They’re very smart and I think they can be of great use,” he said.<\/p>\n Jessica Tillipman, a government procurement law expert, said the federal approach has become inconsistent. “They’ve now adopted completely inconsistent positions across the government,” she told<\/a> The Hill. <\/p>\n She added, “One agency continues to dig its heels\u2026 and the rest is trying to actually work in reality.”<\/p>\n US Split Over Anthropic AI Risk And Access<\/p>\n Earlier, U.S. officials remained divided over Anthropic’s AI systems, with the Pentagon labeling its Claude models a supply chain risk<\/a> and considering phasing them out across agencies.<\/p>\n Pentagon tech chief Emil Michael said Anthropic’s Mythos model could help identify cybersecurity vulnerabilities but described it as a “separate national security moment” requiring broader government safeguards.<\/p>\n White House officials were drafting an AI policy for national security agencies<\/a> promoting the use of multiple AI providers and exploring ways to allow controlled access to Mythos despite the Pentagon’s concerns.<\/p>\n