GTG-1002: Anatomy of the First Autonomous Cyber Attack in History
In mid-September 2025, something unprecedented happened in the world of cybersecurity. For the first time in documented history, a large-scale cyber espionage operation was executed almost entirely by an artificial intelligence system, with human intervention reduced to a minimum. Not an assistant, not a copilot: the AI was the pilot. Anthropic, the company that develops Claude, released a detailed report on this operation, attributed with "high confidence" to a Chinese state-sponsored group designated GTG-1002. It's a watershed moment in the history of information security, and it deserves to be thoroughly understood in all its facets.





