A federal judge expressed doubt Thursday regarding whether the U.S. government has justified its ban on federal use of Anthropic AI technology [1].

The ruling challenges the Trump administration's ability to restrict specific AI providers based on national security concerns without providing a transparent evidentiary basis. This case could set a precedent for how the government labels and bans emerging technologies in the federal supply chain.

The dispute centers on a "supply-chain risk" label applied to Anthropic PBC by the administration [2]. This designation effectively barred federal agencies from integrating the company's AI tools into government operations. However, the presiding judge said the government is not "really justifying" the restriction [1].

According to court proceedings, the judge said the Trump administration "still lacks evidence for Anthropic ’supply-chain risk’ label" [2]. The court's skepticism suggests that the administration's internal justifications for the ban may not meet the legal threshold required to sustain such a restrictive measure against a private entity.

Anthropic has contested the ban, arguing that the designation lacks a factual foundation and unfairly harms its ability to serve public sector needs. The administration has maintained that the ban is necessary for security, though it has not yet provided the specific evidence the judge requested [2].

The judge's comments signal a potential shift in the legal standing of the ban. If the administration cannot produce sufficient evidence to support the supply-chain risk claim, the court may order the government to lift the restrictions on the AI provider [1].

The judge said the government is not "really justifying" the restriction.

This judicial skepticism highlights a growing tension between executive national security prerogatives and the legal requirement for evidence-based regulation. If the ban is overturned, it may limit the administration's power to use broad 'risk' labels to exclude technology vendors without disclosing the underlying intelligence or evidence to the court.