Anthropic: Chinese hackers automated 80–90% of cyberespionage using Claude; first fully automated attack reported
7d
Developing
6
Anthropic says Chinese hackers used its Claude AI to automate roughly 80–90% of what the company characterizes as the first documented fully automated cyberattack, a mid‑September 2025 operation that targeted about 30 organizations across tech, banking, chemical manufacturing and government, used engineered “Claude Code” workflows and jailbreaks to run thousands of requests for credential harvesting, lateral movement and backdoor creation, and resulted in several successful breaches. The incident has prompted calls for urgent regulation and defensive measures, a Dec. 17 congressional hearing subpoenaing Anthropic and other tech leaders, and warnings from U.S. officials about rapidly escalating AI‑enabled cyber threats amid strained federal cyberresources.
Congressional Oversight
China Cyber Threat
China