AI Agents Launch Unprecedented Cyberattack on Taiwan

In a significant escalation of cyber warfare, suspected Chinese hackers have utilized advanced AI tools to launch an autonomous cyberattack targeting Taiwan's government websites. This unprecedented breach, identified by Israeli AI company Dream, showcased the evolving capabilities of AI in cyber operations.

The sophisticated attack deployed up to eight autonomous AI agents simultaneously, which mapped 21 government systems, identified vulnerabilities, and dynamically adjusted their strategies when encountering defenses. The operation, which occurred over four days in early July, successfully compromised at least 85 government user accounts.

During the breach, over 2,500 personnel records were exfiltrated. The AI tool's adaptable nature allowed it to pivot and expand its targets, eventually affecting Taiwan's nuclear safety agency and at least seven energy companies, highlighting a new phase of AI-driven cyber threats.

Researchers suspect a China-linked group due to internal communications in Simplified Chinese, though Taiwan's Ministry of Digital Affairs has not officially commented, citing confidentiality. The attack demonstrates how AI agents can automate complex hacking tasks, posing dual challenges of automated attacks and AI vulnerabilities for network security defenses.