AI tools are making cyberattacks easier by automating vulnerability searches and attack procedures, raising concerns that government-funded research institutes could become targets even when attackers lack advanced expertise. At a parliamentary audit, lawmakers said only 1,654 of 19,041 servers across 23 institutes under the science ministry had antivirus software installed, a rate of 8.7 percent.
Audit findings also cited unauthorized removal of research materials and malware infections on important servers, while separate information-security inspections had rated some institutions favorably. The AI Safety Research Institute under ETRI reportedly has six practitioners assessing cybersecurity risks in AI models, prompting calls for continuous monitoring, attack-based vulnerability testing and AI-assisted defensive systems.