World-first autonomous ‘end-to-end’ AI attack against Taiwan tied to Chinese hackers — and the scariest part is that it was fully open source
China implicated in Taiwan attack through written documentation recovered from the attack.
In a groundbreaking cyberattack, China unleashed a fully autonomous vulnerability hunting assault on Taiwan, targeting government accounts, personnel records, and the nuclear safety agency. This attack marked the first of its kind to utilize eight open-source AI models, enabling the hackers to independently conduct reconnaissance, intrusion, and chain vulnerabilities while adapting to blockades.
The breach, exposed by The Financial Times on August 12, 2026, lasted four days and resulted in the compromise of 85 government accounts and theft of over 2,500 personnel records. Dream, an Israeli AI and cyberdefense company, first identified the breach through a 160MB online archive of 1,395 files. The hackers employed Hermes and OpenClaw, two open-source AI agents, to craft a multi-agent intrusion system.
The use of AI models and the open-source tools elevated the attack's complexity and efficiency. Experts warn that such autonomous AI attacks could pose a significant threat to global cybersecurity, urging governments to adopt stricter security measures.
Written by urgent.news from TechRadar's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
