Yesterday, a federal judge presiding over California's Northern District granted a preliminary injunction, preventing the enforcement of the US government's supply chain risk designation for Anthropic for now. While its analysis of the facts are not final, the judge found that because Anthropic's case is strong and because the government's actions have the potential to immediately damage them, the court should protect the status quo until the arguments of the case have been settled.

I predicted that Anthropic would likely prevail on its first amendment claims and it appears that the judge also found the argument convincing:

Here, Anthropic has shown a likelihood of success on its First Amendment claim. The
record shows that Defendants’ conduct appears to be driven not by a desire to maintain operational control when using AI in the military but by a desire to make an example of Anthropic for its public stance on the weighty issues at stake in the contracting dispute. Although Anthropic had always applied the usage policies in question to Claude Gov, it had been repeatedly lauded as a partner and passed lengthy national security vetting processes. Only when Anthropic went public with its concerns about DoW’s contracting position did Defendants set out to publicly punish Anthropic for its “ideology” and “rhetoric,” as well as its “arrogance” for being unwilling to compromise those beliefs. At that point, Defendants announced a plan to cripple Anthropic: to blacklist it from doing business with any company that services the U.S. military, to permanently cut off its ability to work with the federal government, and to brand it an adversary that could sabotage DoW and that posed a supply chain risk. Those actions go far beyond what would be necessary to address DoW’s ostensible concern about having complete operational control when using AI. This appears to be classic First Amendment retaliation.

In first amendment retaliation cases, one has to prove that the government's primary motivation for action is their exercise of protected speech. The legal term is but-for causation or demonstrating a "nexus." Others may know this as counterfactual analysis: if Anthropic had not engaged in this protected speech, would the government had acted similarly? Anthropic argued, and the judge agreed, that the government acted due to its public statements and advocacy.

In this case, the public outcry from AI Safety communities and developers appeared to, in part, convince the judge. Third parties may file amici curiae or "friend of the court" filings, if they have a strong interest in the outcome of the case. For a laugh, read the Onion's Supreme Court amicus filing on an unrelated first amendment case. The court received numerous amicus briefs warning the court about the chilling effects of the government's actions will have on public discourse:

Several amicus briefs support this conclusion. A group of 37
individuals working on AI technology assert that the Challenged Actions “chill professional debate on the benefits and risks of frontier AI systems and various ways that risks can be addressed to optimize the technology’s deployment.” (Dkt. No. 24-1 at 8.) An industry group of “values-led investors” warns that the Challenged Actions chill speech necessary to allow them to direct their investments to support the “principles and values” they care about.

Anthropic's case falls pretty squarely into textbook first amendment retaliation. Despite the government's broad powers over national security and administrative policy, courts generally frown upon the employment of government offices to punish individuals purely for engaging in protected speech.

The court also analyzed Anthropic's Administrative Procedure Act (APA) claims. As I noted, Anthropic both claimed the designation was both beyond the powers of the government and "arbitrary and capricious," or lacking a rational basis required for agency decisions. While the federal government has broad powers to enact change, federal agencies are required to follow many procedures before making a policy determination, as required by law. In Anthropic's case, a supply chain risk designation required that no other reasonable and less intrusive measures to protect national security were available. The judge held that its likely no other courses of action were considered at all. The judge was particularly sharp:

The timing of the administrative record raises an inference that it was generated after the fact to justify the foreordained conclusion ordered in Secretary Hegseth’s mandate to “designate Anthropic a Supply-Chain Risk to National Security” on February 27.

While I was surprised at the cultural backlash against the DoW and in support of Anthropic, I am not surprised by the court's unwillingness to support the government's naked attempt at corporate crucifixion. Executive fiat is not capable of overruling legislatively imposed due process nor first amendment rights.