Anthropic said no to killer robots. The Pentagon called that a national security risk.

Anthropic said no to killer robots. The Pentagon called that a national security risk. Say it out loud and it sounds like a joke. Nobody in Washington is laughing. On Friday, a federal appeals court in Washington voted 2-1 to let the Pentagon keep calling Anthropic — the company behind Claude — a national security supply-chain risk. The reason, once you boil it down: Anthropic refused to let the military use its AI for autonomous weapons and mass surveillance. That's the whole story. The safety lab said no to killer robots, and the government decided the refusal was the threat. Defense Secretary Pete Hegseth slapped the label on Anthropic earlier this year and ordered Claude ripped out of military and federal systems. Anthropic says the designation is costing it billions in lost contracts and staining the company's name right before an anticipated IPO. The courts are split. A federal judge in San Francisco already struck down a parallel designation under a different law, finding the administration had unlawfully retaliated against Anthropic for its views on AI safety. Friday's ruling goes the other way. When two courts disagree this hard, there's only one door left: the Supreme Court. Here's the precedent everyone should be sweating. If refusing to weaponize your AI is what makes you the security risk, the industry just learned what "cooperation" means. And it's not a choice anyone should envy. Sources: Reuters, "US appeals court declines to block Pentagon's blacklisting of Anthropic" (Sept 25, 2026); The Wall Street Journal, "U.S. Appeals Court Upholds Pentagon Labeling Anthropic a Security Risk" (Sept 25, 2026).