A federal appeals court ruled 2-1 in favor of the U.S. Department of Defense, upholding its decision to blacklist AI developer Anthropic from military contracts following a feud over safety guardrails.
Key Takeaways
- A federal appeals court ruled 2-1 to uphold the Pentagon's decision blacklisting Anthropic from military contracts.
- The dispute escalated after Defense Secretary Pete Hegseth demanded Anthropic remove safeguards blocking its Claude AI from autonomous lethal weapons and domestic surveillance.
- The decision directly conflicts with an August ruling by California federal Judge Rita Lin, who deemed a parallel ban unconstitutional.
Background of the Pentagon Dispute
The legal battle traces back to February, when Defense Secretary Pete Hegseth demanded that Anthropic strip safety guardrails preventing its AI model, Claude, from being used in fully autonomous lethal weapons and mass domestic surveillance. Anthropic publicly refused on February 26, hours before Hegseth's deadline. In response, Hegseth formally classified Anthropic as a national security supply chain risk on March 3, canceling existing military contracts and barring other defense contractors from using its software.
According to court documents, defense officials cited an incident where an Anthropic executive questioned a contractor's use of Claude during a specific military operation—identified in case records as the January 3 mission that captured Venezuelan President Nicolás Maduro. Writing for the majority alongside Judge Neomi Rao, Judge Gregory Katsas stated the Pentagon had "ample support" for its conclusion that Claude's behavior was too unpredictable for military reliance, while Judge Karen LeCraft Henderson dissented.
Conflicting Rulings and Legal Uncertainty
Friday's decision creates a sharp divide within the federal judiciary. On August 28, federal Judge Rita Lin in California ruled a broader parallel Trump administration ban unconstitutional under the First Amendment and Fifth Amendment, characterizing the blacklist as unlawful retaliation designed to make a public example of the company.
Responding to the appellate decision, an Anthropic spokesperson told CNBC, "We respectfully disagree with the court's decision," adding that "another federal court has already held the government's parallel designation unlawful." As public institutions examine broader AI safety and government integration concerns, it remains unclear whether Anthropic will request a full circuit rehearing or appeal to the U.S. Supreme Court.
Why It Matters
This ruling underscores a fundamental tension between defense agencies seeking unrestricted AI deployment and commercial technology companies attempting to enforce ethical boundaries. With opposing federal court decisions now active, defense vendors face significant legal ambiguity regarding software integrations. The case sets a crucial legal precedent on whether the federal government can blacklist private contractors for maintaining safety restrictions on dual-use commercial technologies.
