Anthropic report: Claude AI used by adversaries for drone, spying, virus research

Anthropic's threat report details five ways its Claude model has been exploited by U.S. adversaries for developing kamikaze drones, hunting dissidents, and conducting dangerous virus research. This marks a concrete escalation in AI-enabled asymmetric threats, lowering the barrier for small groups to conduct operations previously requiring state-level resources. The report underscores immediate national security risks from current AI models, distinct from existential long-term concerns.