The court has ruled that Anthropic must be removed from the Pentagon's blacklist.

A U.S. federal court in California has ruled that the Pentagon’s decision to blacklist Anthropic as a “supply chain risk” was unlawful. The judgment, issued by District Court Judge Rita F. Lin, found that the classification violated constitutional protections and was motivated more by political considerations than legitimate national security concerns. The case, which began in early 2023, centered around Anthropic’s refusal to permit the unrestricted military use of its artificial intelligence models. The ruling does not immediately remove Anthropic from the list but sets a legal precedent that challenges the Pentagon’s authority to impose such restrictions. The dispute originated in February when then-President Donald Trump and Defense Secretary Pete Hegseth accused Anthropic of posing a threat to national security after the company declined to allow the military to use its AI technology without limitations. Anthropic’s CEO, Dario Amodei, stated in a blog post that AI systems lacked the reliability required for fully autonomous weapons and expressed concerns about their potential use in mass surveillance. This stance drew sharp criticism from Trump, who labeled Anthropic a “radical left, woke company” and called for a ban on its products within federal agencies. Hegseth followed up by urging businesses to sever ties with Anthropic if they wished to engage in defense contracts. In March, the Pentagon formally classified Anthropic as a “supply chain risk,” a rare move targeting a U.S.-based company. The designation effectively barred Anthropic from receiving defense contracts and restricted its ability to supply other firms that might contract with the military. Anthropic responded by arguing that the action was part of an “unlawful campaign of retaliation” and warned that it would harm the company’s broader business interests beyond defense-related dealings. The company also pointed out that the Pentagon had previously engaged in discussions with it regarding the use of its AI model, Mythos, contradicting the claim that it posed a serious threat. Judge Lin’s 59-page ruling emphasized that the government’s justification for the blacklisting, relying heavily on vague national security claims, was insufficient to justify the severe economic repercussions imposed on Anthropic. She noted that there was no evidence that Anthropic had inserted any backdoors into its AI models, as the Pentagon had alleged. Additionally, she highlighted inconsistencies in the government’s position, including recent conversations with Anthropic and statements from Hegseth himself, who had previously described the company as essential to national security. Lin concluded that the government’s actions appeared to be driven by a desire to publicly shame Anthropic rather than respond to actual threats. The ruling comes amid growing scrutiny of how the government regulates emerging technologies. Earlier in the legal process, Lin had temporarily blocked the enforcement of Trump’s order directing federal agencies to cease using Anthropic’s AI tools. She also expressed concern that the government’s approach conflicted with the First Amendment, which protects freedom of speech and expression. The judge’s decision has been widely interpreted as a rebuke of efforts to suppress dissent through regulatory means, particularly in the context of AI development. The Pentagon retains the right to appeal the decision, though it has not yet indicated whether it will do so. Meanwhile, Anthropic is preparing for an initial public offering (IPO) that could raise billions of dollars, signaling confidence in its future despite the ongoing legal battle. The outcome of the case may influence how the government handles similar disputes involving private companies and national security concerns in the rapidly evolving field of artificial intelligence.