Anthropic publishes September 2026 threat intelligence report on Claude misuse

Anthropic detailed and disrupted multiple malicious uses of Claude between December 2025 and August 2026 spanning cyber operations, state surveillance, influence campaigns, conventional weapons, biological research and illicit distillation. Notable cases included a Russia-linked espionage campaign and large-scale distillation attributed to Chinese firms. The report underscores rising autonomy in adversarial AI workflows and the need for shared defensive measures across labs and governments.