China-linked hackers used AI agents to breach Taiwan government systems

Suspected Chinese hackers deployed autonomous artificial intelligence agents to breach Taiwanese government websites in what researchers are calling an unprecedented display of AI-driven cyber operations, according to an investigation first reported by the Financial Times on August 11.ft

Over four days at the start of July, the AI-powered hacking tool simultaneously deployed up to eight autonomous agents that mapped 21 government systems, identified vulnerabilities, and adapted tactics when blocked by defenders, according to researchers at Dream, an Israeli AI and cybersecurity firm that uncovered the intrusion.slguardian+1

An End-to-End Autonomous Attack

The attack compromised at least 85 government user accounts and extracted more than 2,500 personnel records before expanding to Taiwan's nuclear safety agency and at least seven energy companies, Dream's research found.fudzilla+1

Dream's chief strategy officer Amir Becker, a former head of cyber operations for Israel's elite signals intelligence Unit 8200, said he had never seen such an "end-to-end autonomous attack" against a government target.slguardian+1

The hacking tool was built on two open-source AI agent frameworks, Hermes and OpenClaw, which allow AI models to perform tasks autonomously. Researchers found evidence of the breach in a 160MB online archive containing 1,395 files. The tool's safety guardrails were bypassed by framing the activity as authorized penetration testing.nationaltechnology+2

Dream has not attributed the operation to a specific group but noted that Simplified Chinese in internal communications indicated a high probability the operator was connected to China. The stolen data was written in Traditional Chinese, commonly used on Taiwanese government websites.slguardian+1

A New Dimension to an Escalating Threat

The breach adds a new layer to an already intense cyber campaign against Taiwan. The island's National Security Bureau reported in January that it faced a daily average of 2.6 million Chinese cyberattacks in 2025, more than double the figure from 2023.globaltaiwan+1

Taiwan's Ministry of Digital Affairs declined to comment on the specific incident but acknowledged that AI-enabled attacks have transformed the nature of cybersecurity threats, noting that AI agents create a "dual challenge" because attacks can be automated while the agents themselves can become vulnerabilities.slguardian

The incident follows reports from Anthropic last November that Chinese state-sponsored hackers had manipulated its Claude tool to attempt attacks on 30 international organizations, though with limited success. "This must be the basic assumption of every government around the globe," Becker said of the threat of permanent automated cyber siege.fudzilla+1