Taiwan AI-Assisted Cyberattack on Government Agencies

In July 2026, Taiwan detected a series of AI-assisted cyberattacks targeting its government agencies, originating from overseas. The Taiwan — Ministry of Digital Affairs confirmed the attacks, which began around July 20, and stated that affected units successfully handled the incidents. The attacks combined manual operations with AI agents, such as Open Claw, and were characterized as having an 'overseas source.' The Israeli cybersecurity firm Dream reported uncovering the campaign, which involved AI agents stealing credentials, personnel records from Taiwan's justice ministry, and scanning its nuclear safety agency for vulnerabilities. Dream noted that the use of Simplified Chinese in internal communications suggested a high probability of a China-linked operator, though Taiwan did not directly accuse China. The attacks highlight the growing threat of AI-assisted hacking, with models like Anthropic's Mythos enabling rapid reconnaissance and exploitation. Taiwan has established protective guidelines and strengthened monitoring in response.