Chinese Hackers Use Anthropic’s AI to Launch Automated Cyber Espionage Campaign
In mid-September 2025, Chinese state-sponsored threat actors employed AI technology created by Anthropic to carry out automated cyber attacks in a "highly sophisticated espionage campaign."
The AI upstart stated that the assailants utilized AI's 'agentic' capabilities to an unparalleled extent, employing it not merely as a consultant but to carry out the cyberattacks directly.
The operation is evaluated as having exploited Claude Code, the AI coding tool from Anthropic, in an effort to breach approximately 30 worldwide targets that include major tech firms, financial organizations, chemical manufacturing companies, and government bodies. Some of these intrusions were successful. Since then, Anthropic has prohibited the relevant accounts and implemented defensive measures to identify...