AI cybersecurity group presents 13 papers at Korean conference
The special session in Gangneung on 23 October covered AI security controls, LOLBins detection, generative AI privacy and AI agent permissions.
The AI Cybersecurity Research Group, jointly run by the International Cybersecurity Certification Association (CSC) and the Korean Institute of Internet Information Sciences, held a special session with 13 paper presentations at the institute's 2026 autumn academic conference in Gangneung, Gangwon Province, on 23 October, according to dailysecu.com.
The session covered AI security controls in cloud environments, detection of attacks abusing legitimate executables, generative AI privacy, AI agent permission management and data-leak prevention.
CSC president Kong Byeong-cheol presented research on extending AI security controls in ISO/IEC 27017 cloud service areas and covered ISO/IEC 27090:2026, described as enacted on 15 October. Kong noted that ISO/IEC 27000:2026 is the sixth revision and ISO/IEC 27017:2026 is the second revision.
Director Yeo Dong-gyun presented an AI model for detecting LOLBins attacks based on process behavior sequences, while Director Lee Jun-hwa presented a study on supplementing ISMS-P privacy impact assessment for generative AI use. Director Nam Hyeok-woo presented a five-layer security control model against excessive permissions and indirect prompt injection in AI agent environments.
Director Lim Beom-seok presented research on data movement security threats and leak prevention for N2SF-based generative AI services, and senior researcher Park Tae-gyun presented research using RAG-based LLMs to analyze false positives in web vulnerability diagnosis and support diagnoser decisions.
Quick answers
Where and when did the AI Cybersecurity Research Group hold its special session?
The session took place on 23 October at Saint John's Hotel in Gangneung, Gangwon Province, during the Korean Institute of Internet Information Sciences 2026 autumn academic conference.
How many papers were presented at the session?
Thirteen papers were presented, covering topics including AI security controls in cloud, LOLBins attack detection and generative AI privacy.