China-Linked Hackers Conduct First Autonomous Cyberattack on Taiwan Using AI

China-Linked Hackers Conduct First Autonomous Cyberattack on Taiwan Using AI

First seen 12 Aug 2026, 20:38 UTC Feeds.4SysopsRss.Slashdot 71% similarity 75.0

Article Content

Browse articles
ThreatCluster

Suspected China-linked hackers executed an end-to-end autonomous cyberattack on Taiwanese government systems, utilizing open-source AI-agent frameworks Hermes and OpenClaw. The attack, which lasted four days in July, compromised at least 85 accounts and resulted in the theft of over 2,500 personnel records. The autonomous system was capable of adapting its attack paths without human intervention, continuously assessing and reprioritizing its strategies. The attack targeted various government entities, including Taiwan's nuclear safety agency and several energy companies. Researchers from Israeli cybersecurity firm Dream discovered evidence of the attack in a 160MB archive containing 1,395 files. The attackers presented the intrusion as an authorized penetration test to bypass safeguards. The use of easily accessible tools raises concerns about the potential for similar attacks by other threat actors.

Key Points: • First known instance of an autonomous cyberattack using AI against a government. • Attack compromised 85 accounts and stole over 2,500 personnel records. • Utilized open-source AI frameworks Hermes and OpenClaw for attack execution.

ThreatCluster AI How this analysis works

Timeline

2026-07-01
Autonomous cyberattack on Taiwan began
Suspected China-linked hackers launched a four-day autonomous attack on Taiwanese government systems, compromising multiple accounts.
Rss.Slashdot
2026-07-04
Attack concluded
The autonomous cyberattack ended after four days, resulting in significant data theft from various government agencies.
Rss.Slashdot
2026-08-12
Details of the attack revealed
Israeli cybersecurity firm Dream published findings on the attack, highlighting the use of AI frameworks and the scale of the breach.
Feeds.4Sysops

Community

Browse all →

Tracked Entities in This Story