In November 2025, Anthropic published a threat disclosure about a Chinese state-linked actor that had automated most of an active intrusion campaign using