A federal appeals court has issued a ruling upholding the Trump administration's decision to blacklist the artificial intelligence company Anthropic. The legal dispute centered on the company's refusal to enable specific features within its Claude AI model, which government officials argued were necessary for national security and military readiness.

The court’s decision focused on the operational implications of AI software used by the government. Judges concluded that the use of what they described as "overly constrained" AI models could lead to failures in military operations. By restricting the functionality of its software, the court reasoned, Anthropic created a security vulnerability that justified the administration's decision to place the company on a restricted entity list.

Anthropic had previously argued that its restrictions were intended to ensure safety and prevent the misuse of its technology. However, the court prioritized the government's assessment of military necessity over the company's internal safety protocols. This ruling establishes a significant precedent regarding the extent to which the federal government can mandate specific technical capabilities from private AI developers when those tools are integrated into government or defense systems. The decision effectively bars federal agencies and contractors from utilizing Anthropic’s services until the company complies with the government's feature requirements.