Chinese Hackers Conduct Autonomous AI Attack on Taiwan's Government Systems

Chinese hackers used publicly available artificial intelligence tools to attack Taiwan's government systems. The attack lasted four days in early July, involving up to eight AI agents that independently explored the network and searched for vulnerabilities. The attackers compromised at least 85 government accounts and obtained over 2,500 records of employees. The attack spread to Taiwan's Nuclear Safety Agency and at least seven energy companies. The tool utilized two open AI agent systems—Hermes and OpenClaw—that allow tasks to be performed autonomously. Researchers consider the incident one of the first examples of a fully autonomous cyberattack. Internal communications related to the attack used simplified Chinese characters, indicating a possible connection of the operators to China. A representative from Taiwan's Ministry of Digital Affairs did not confirm the specific breach but noted that the use of AI is changing the nature of cyberattacks.