Taiwan Detects AI-Assisted Cyberattacks
Taiwan’s Ministry of Digital Affairs says that government agencies were targeted last month by artificial intelligence-assisted cyberattacks originating from overseas, but that the affected bodies successfully handled the incident after it was detected and investigated.
In a statement, the ministry reported that its cybersecurity monitoring units identified an abnormal attack on government systems in July. Beginning on 20 July, the National Institute of Cyber Security issued a series of warning alerts while investigators examined the activity. The inquiry found clear characteristics of an overseas source, with the attackers using a hybrid approach that combined manual operations and AI agent-assisted techniques, including tools such as Open Claw.
“The relevant attack sources, methods, and scope of impact have all been fully investigated, and the affected units have successively completed their handling,” the ministry said. In response to this new form of AI-enabled threat, the government has established protective guidelines and strengthened system monitoring across agencies to detect and block similar attacks at an early stage.
The official confirmation follows reporting by the Financial Times and research from the Israeli cybersecurity firm Dream, which described an AI-driven campaign that used publicly available AI agents to operate in a near-autonomous manner. According to the researchers, the attackers deployed multiple AI agents that mapped government systems, researched vulnerabilities, adapted tactics when blocked, compromised dozens of user accounts and extracted personnel records before expanding their activity toward Taiwan’s nuclear safety agency and several energy companies. The operation unfolded over several days in early July.
Taiwan has long said it faces frequent cyber operations that form part of what it describes as hybrid pressure from China, alongside military activity and disinformation campaigns. Beijing claims sovereignty over the democratically governed island and has not commented on the latest incident. The Ministry of Digital Affairs statement did not attribute the attacks to any specific country.
Autonomous or semi-autonomous agents are known to accelerate reconnaissance, vulnerability discovery and adaptation, potentially increasing the speed and scale of attacks against government and critical infrastructure targets. Taiwanese authorities say they have already adjusted defensive measures in light of the new techniques observed last month.
