Suspected China-linked hackers used AI to run the first-ever end-to-end autonomous cyberattack on Taiwan's government, Israeli firm says — open-source-built tool continuously devised effective hack strategies in real-time
Suspected China-linked hackers used autonomous AI agents to breach Taiwanese government systems, compromising 85 accounts and stealing 2,500+ records.
Chinese-linked hackers employed AI to orchestrate the world's inaugural end-to-end autonomous cyberattack on Taiwan's government, as per an August 12 Financial Times report citing Israeli cybersecurity firm Dream. The breach resulted in compromised accounts and theft of over 2,500 personnel records. The attackers constructed an autonomous hacking platform utilizing open-source AI-agent frameworks, Hermes and OpenClaw, enabling agents to map networks, discover vulnerabilities, intrude, and modify tactics when a path faltered.
The campaign lasted four days in early July and employed up to eight agents concurrently. Dream discovered the intrusion in a 160MB online collection containing 1,395 files, indicating the tool's foundation on two accessible AI agents. The assailants shifted the intrusion as a penetration test rather than a genuine assault, sidestepping inherent safeguards.
Notably, the tool autonomously devised attacks, adapting to evolving circumstances. While Dream refrained from attributing the campaign to a specific group or nation, Chinese-language internal communications and Traditional Chinese data suggest a possible Chinese connection. The incident underscores growing worries in the cybersecurity and AI sectors about autonomous AI models capable of executing complex cyberattacks.
Written by urgent.news from Tom's Hardware's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
