Israeli AI security firm Dream reported that the cyber campaign unfolded over four days in July, during which the autonomous system mapped 21 government systems, compromised 85 user accounts and extracted roughly 2,500 personnel records. The operation further extended to Taiwan’s nuclear safety agency, multiple government IT suppliers and at least seven companies operating in the energy sector. Dream added that the AI platform directed as many as eight specialized agents in carrying out reconnaissance, credential attacks, vulnerability exploitation and dynamic path planning.
Taiwan’s Ministry of Digital Affairs stated that its probe identified clear evidence of overseas origins for the attacks, which combined traditional hacking tools with AI agents in a hybrid model. The ministry stopped short of identifying any responsible party while noting that the incident illustrated the growing role of autonomous technologies in cyber operations. Officials indicated that such events call for enhanced defensive measures along with updated legal and policy frameworks to counter AI-driven threats.
According to Dream, the system demonstrated an ability to adapt its approach when initial attempts were thwarted, proceeding without the need for continuous human oversight at each stage. Internal documents associated with the intrusion used distinctive character forms consistent with Simplified Chinese, pointing to possible links with overseas actors although no formal attribution was made. The firm recovered a complete operational workspace from a 160-megabyte online archive that enabled reconstruction of the full attack sequence.
Kenny Huang, chairman of the Taiwan Network Information Center, described the incident as the first fully automated cyberattack against a government of which he was aware. The event has amplified global concerns regarding the potential for capable AI systems to reduce the costs and accelerate the pace of malicious cyber activities. In statements following the disclosure, Taiwanese officials highlighted the necessity of stronger capabilities to address these evolving challenges.
Taiwan’s National Security Bureau data showed that cyberattacks on the island’s key infrastructure averaged 2.6 million per day in 2025, a 6 percent rise from the previous year. Many of these incidents were linked to actors in China and at times coordinated with military exercises as part of hybrid warfare tactics, the bureau reported. The sharp increase, which exceeded 100 percent in some metrics since 2023, has placed sustained pressure on Taiwan’s digital defenses.
The Ministry of Digital Affairs has outlined the Phase Seven National Cyber Security Development Program covering 2025 through 2028 to tackle such risks through greater use of AI for defense, talent development and collaboration with the private sector. This program focuses on whole-of-society resilience, protection of critical infrastructure and applications of emerging technologies. It forms part of a broader strategy that treats cybersecurity as a core element of national security.
Добавить комментарий