Autonomous AI Hackers Breach 85 Taiwan Accounts, Steal 2,500 Records
Updated
Updated · CNN · Aug 13
Autonomous AI Hackers Breach 85 Taiwan Accounts, Steal 2,500 Records
3 articles · Updated · CNN · Aug 13
Summary
Over four days in July, autonomous AI agents mapped 21 Taiwan government systems, cracked 85 user accounts and extracted 2,500 personnel records in what experts call the first disclosed fully automated attack on government agencies.
Dream said the hackers used open-source agents including OpenClaw to automate reconnaissance, credential attacks and next-step decisions, with up to eight AI agents adapting in real time without human intervention.
Taiwan’s digital affairs ministry said the campaign also hit the nuclear safety agency, government IT vendors and at least seven energy companies, exploiting secondary systems to scale the intrusion faster and more cheaply.
Simplified Chinese in internal documents points to a high probability of China links, experts said, though Taiwan and Dream stopped short of attribution and China’s foreign ministry said it was unfamiliar with the case.
The breach lands as Taiwan says it faced 2.6 million cyberattacks a day from China last year, underscoring warnings that AI is cutting attack costs faster than governments are building defenses.