Anthropic Identifies AI-Driven Cyber Espionage by Chinese Group

Anthropic has uncovered a sophisticated cyber espionage campaign orchestrated entirely by artificial intelligence (AI), targeting around thirty global organizations. The attack, attributed to a suspected Chinese state-sponsored group, was first identified in mid-September 2025 and involved the manipulation of Anthropic’s Claude Code tool. While the attack was not financially successful in its initial stages, it marks a significant advancement in AI’s role in cyber operations.