< Back to all clusters
[TECHNOLOGY] · Taiwan · 5 sources

started · updated

Taiwan government targeted by first autonomous AI-driven cyberattack

Israeli cybersecurity firm Dream has documented what is described as the first end-to-end autonomous cyberattack against a government target. The operation, which targeted Taiwan, utilized open-source AI agents, specifically Hermes and OpenClaw, to function like a coordinated hacking team.

During a four-day period in early July, the attack deployed up to eight autonomous agents simultaneously. These agents mapped 21 government systems, identified vulnerabilities, and independently altered tactics when blocked by defenders. The breach resulted in the compromise of at least 85 government user accounts and the theft of over 2,500 personnel records.

The campaign eventually expanded beyond general government systems to target Taiwan’s nuclear safety agency and at least seven energy companies. While Dream did not officially attribute the attack to a specific group, researchers noted that internal communications used Simplified Chinese, suggesting a potential connection to China. The incident highlights a shift toward AI-driven warfare where tools can independently choose targets and rank techniques with minimal human oversight.

Entities

Amir Becker · China · Dream · Hermès · Taiwan · Taiwan government

Claims

What the coverage asserts, and how well corroborated each claim is across sources.