Taiwan Flags AI-Assisted Cyberattacks Targeting Government Agencies

Taiwan detected a wave of AI-assisted cyberattacks targeting government agencies in July, with authorities tracing the malicious activity to overseas sources, the Ministry of Digital Affairs said Thursday. Officials said affected agencies successfully contained and handled the attacks.

Taiwan has increasingly warned about cyber threats amid heightened tensions with China, which Taipei accuses of conducting “hybrid warfare” through military activity, disinformation and hacking campaigns. Taiwan’s National Security Bureau previously reported that cyberattacks against critical infrastructure, including hospitals and banks, averaged 2.63 million attempts per day in 2025, up 6% from the previous year.

The Ministry of Digital Affairs said cybersecurity monitoring teams identified abnormal activity in July. Beginning July 20, the National Institute of Cyber Security issued multiple alerts while investigating the attacks.

According to the ministry, investigators found characteristics indicating an overseas origin. Hackers reportedly used a hybrid strategy combining manual operations with AI agent-assisted cyberattacks, including tools such as Open Claw. Authorities said the sources, techniques and impact of the campaign were investigated and affected agencies completed necessary remediation.

Taiwan has since introduced new cybersecurity protection guidelines and strengthened government system monitoring to detect and block similar AI-powered attacks earlier. The ministry did not directly attribute the campaign to China.

The announcement followed a report from Israeli cybersecurity company Dream, which said it uncovered an AI-driven hacking operation targeting an unnamed Asian government. Dream reported that coordinated AI agents extracted passwords, obtained personnel information and scanned government systems for security vulnerabilities during a four-day campaign. The Financial Times identified the targeted organizations as Taiwanese agencies.

Cybersecurity researchers say the incident highlights the growing capabilities of AI-powered hacking tools. Newer AI models can accelerate reconnaissance, vulnerability discovery and exploitation, potentially allowing attackers to conduct sophisticated campaigns more quickly.

However, security experts caution that AI agents are not necessarily operating independently. Cris Thomas, a security advocate at Semgrep, noted that human operators still determine targets, objectives and instructions.

The incident underscores Taiwan’s growing challenge of defending government networks as artificial intelligence becomes an increasingly important tool in global cyber warfare.