Anthropic Says Iran, Russia Used Claude for Weapons Research

Anthropic publicly accused state actors from Iran and Russia of using Claude to assist weapons research. This is the first time a major AI lab has named specific countries, directly linking model misuse to geopolitical adversaries. The post doesn't disclose weapon types, which Claude versions were used, or how Anthropic detected and attributed the activity. I'd treat this as a one-sided statement for now and wait for more technical details before assessing the actual harm.

Why it matters: Anthropic's first public accusation of nation-state actors using Claude for weapons research scores high on H and R. But the post lacks weapon type, model version, and detection details, so K is absent — keeping it below 85.