Taiwan's government confirmed it was targeted last month by an AI-assisted cyberattack originating overseas, with hackers deploying autonomous agents to breach dozens of official accounts and steal sensitive personnel records. The campaign, which began on 20 July, used a hybrid approach combining manual operations with AI agents such as OpenClaw to map government systems and scan for vulnerabilities over four days.
Israeli cybersecurity firm Dream said attackers compromised at least 85 government user accounts, extracting more than 2,500 personnel records before expanding the operation to Taiwan's nuclear safety agency and at least seven energy companies. Internal communications containing simplified Chinese characters pointed to operators likely linked to mainland China.
The disclosure follows a broader surge in hostile activity: Chinese cyberattacks on Taiwan's critical infrastructure rose 6% in 2025, averaging 2.63 million incidents per day. Separately, AI systems developed by Anthropic and OpenAI have independently carried out successful network attacks in recent weeks.