A federal judge has found that the Trump administration unlawfully retaliated against AI company Anthropic after the firm publicly refused to remove key safety limits from its Claude large language model for military uses. The ruling could have broad implications for how governments engage with AI vendors — and for any tech firm weighing whether to cooperate with defense contracts. The dispute began when the Pentagon pressed Anthropic to remove usage restrictions and accept a clause allowing “all lawful uses.” Anthropic complied with most requests but maintained two “red lines,” declining to permit uses such as mass surveillance of Americans and lethal autonomous weapons. On February 27, former President Trump instructed all federal agencies to stop using Anthropic’s technology, calling the firm a “RADICAL LEFT, WOKE COMPANY” on Truth Social. Shortly afterward, Hegseth publicly accused Anthropic of “a master class in arrogance and betrayal” and barred military contractors from doing business with it. Key government concession and timeline - The government ultimately abandoned its central security claim — that Anthropic retained backdoor access to deployed models. Officials now concede Anthropic has no such access and that Claude poses no greater risk than other “black box” AI systems. - With that line of argument gone, the government relied on a loss-of-trust claim tied to what a Pentagon memo described as Anthropic’s “increasingly hostile manner through the press.” - Judge Lin found the timing undermined that justification. Days before the blacklisting, Hegseth had suggested invoking the Defense Production Act, which would have made Anthropic essential to national security. The day after the designation, an undersecretary emailed Anthropic saying a contract was “very close.” Lin also noted the government is still discussing work with Anthropic on its new model, Mythos, in sensitive contexts. Court findings and relief Judge Lin concluded the government’s actions amounted to retaliation in violation of the First Amendment, were imposed without the pre-deprivation procedures required by the Fifth Amendment, fell outside the relevant supply-chain statute, and were arbitrary and capricious. “The empty invocation of national security is not a blank check to punish and retaliate against government critics,” she wrote. The court: - Vacated the federal designation and Hegseth’s boycott order, - Entered judgment for agencies that took no action, - Granted a permanent injunction, - Denied the government’s request for a seven-day administrative stay, noting the government had operated under a preliminary injunction since March without identifying any harm. What Anthropic and the government said Anthropic told the court that keeping the restrictions in place would slash its defense-related revenue by 50% to 100% and cut 2026 revenue by billions. The court, however, rejected Anthropic’s separate argument that Trump’s directive exceeded presidential authority. Why this matters to tech and crypto communities The ruling underscores limits on the government’s ability to blacklist or boycott tech vendors for political reasons — and highlights how national-security rhetoric can be used to pressure companies. For AI developers and other platform providers (including those in crypto and decentralized tech), it signals that refusing certain government uses on ethical grounds may be protected speech, and that procurement and supply-chain decisions must follow proper legal and procedural channels. Disclosure: This article was generated using Anthropic’s Claude LLM. Read more AI-generated news on: undefined/news