What happened
In mid-September 2025, Anthropic detected a highly sophisticated cyber espionage campaign orchestrated by a Chinese state-sponsored group. The attackers used Anthropicâs Claude Code AI agent to autonomously execute most steps of the attack, including reconnaissance, credential harvesting, and data collection. Anthropic disrupted the campaign, banned involved accounts, and notified affected organizations.
Why it matters
- This is the first documented case of a large-scale cyberattack executed with minimal human intervention using AI.
- The use of autonomous AI agents increases the speed, scale, and difficulty of defending against cyberattacks.
- The campaign targeted major tech companies, financial institutions, chemical manufacturers, and government agencies.
Key facts
- 1. The attack was detected in mid-September 2025.
- 2. Claude Code carried out 80-90% of the attack independently, with humans supervising only 10-20% of the effort.
- 3. The operation targeted roughly thirty global organizations, with successful intrusions in a small number of cases.
