Taiwan’s Ministry of Digital Affairs reported that overseas hackers targeted government agencies in July in an “abnormal” cyberattack combining manual operations with AI agent-assisted techniques.
According to Reuters, monitoring units detected the threat, prompting the National Institute of Cyber Security to issue warning alerts beginning July 20 while conducting an investigation. The ministry stated that the attack sources, methods, and scope of impact were fully investigated, with affected agencies successfully completing mitigation procedures.
The statement followed a report from Israeli cybersecurity firm Dream, which uncovered an AI-driven hacking campaign targeting an Asian government. Dream’s analysis revealed that a network of AI agents collaborated over four days to extract credentials from officials, steal personnel records from Taiwan’s Justice Ministry, and scan its nuclear safety agency for vulnerabilities. The Financial Times subsequently identified the target agencies as Taiwanese.
The incident highlights a broader shift toward AI-derived cybersecurity threats, where threat actors leverage newly released models, including Anthropic’s Mythos, to automate reconnaissance and exploit system vulnerabilities. Cybersecurity researcher Cris Thomas noted that while AI agents accelerate the speed of such campaigns, operators still direct the targets and objectives rather than the attacks being fully autonomous. In response to the evolving threat, Taiwanese authorities have updated protective guidelines and enhanced inter-agency system monitoring to block future intrusions early.
