AI 국방정보시스템 도입 시 예상되는 AI 고유 보안 취약점 분석 및 보안 가이드라인 연구
A Study on AI-Inherent Security Vulnerabilities and Security Guidelines for AI-Integrated Defense Information Systems
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
국방부는 『국방혁신 4.0』을 통해 군내 AI 도입을 적극적으로 추진하고 있다. 그러나 이는 AI 장점들과 함께, 기존의 취약점과는 다른 AI 고유의 취약점 또한 가져옴을 의미한다. 따라서, AI 고유의 취약점들에 대한 고유의 보안대책 수립이 필수적이다. 본 연구에선, 국방정보시스템에 AI가 도입됨에 따른 AI 고유의 잠재적 취약점에 대한 완화방안들을 살펴보고, 국방정보시스템의 개발공정 단계별 세부 보안활동 가이드라인을 제안한다. 이를 통해, 향후 AI를 적용한 국방정보시스템의 안전한 도입 및 운용에 기여할 수 있을 것으로 기대한다.
Abstract
The Ministry of National Defense has identified "Securing AI-based Core Advanced Power" as a key task within the "Defense Innovation 4.0" initiative and is vigorously promoting the integration of AI within the military. This strategy, while offering new advantages, also introduces AI-inherent vulnerabilities distinct from traditional ones. Consequently, it is imperative to develop security measures specifically designed to address these AI-inherent vulnerabilities in the defense sector. To this end, this study investigates the AI-inherent vulnerabilities that may emerge with the implementation of AI in defense information systems and proposes corresponding mitigation strategies. Furthermore, we outline comprehensive security activities tailored to each stage of the development process of defense information systems. By doing so, we aim to facilitate the safer adoption and operation of AI-integrated defense information systems in the future.
참고문헌 (41)
- "챗GPT 등 생성형 AI 활용 보안 가이드라인" (2023)[URL]
- 국방부, "국방 정보화업무 훈령", 국방부훈령 제2842호, ~ (2023)
- "2023~2037 국방과학기술혁신 기본계획" (2023)[URL]
- "국방혁신 4.0 기본계획" (2024)[URL]
- 국방부, "국방 사이버보안 위험관리 지시", 국방부기타 제15호, ~ (2024)
- "AI 과학기술 강군 육성’을 위한 발돋움, ｢국방 데이터･인공지능 확산주간｣ 행사 개최" (2024)[URL]
- "국방혁신 4.0 핵심‘AI’ 연구로 최정예 육군 건설 박차" (2024)[URL]
- "유럽연합 인공지능법" (2024)[URL]
- "공공부문 초거대 AI 도입 활용을 위한 가이드라인" (2023)[URL]
- 박규도, "AI 기반 국방정보시스템 개발 생명주기 단계별 보안 활동 수행방안", 정보보호학회논문지, 33(3), pp.577~589 (2023)
- "유럽의회, ｢AI법｣ 가결" (2024)[URL]
- 윤정현, "국방 분야 인공지능 기술 도입의 주요 쟁점과 활용 제고 방안", 과학기술정책, (279), pp.1~55 (2021)
- "AI 첨단전력 확보…근무환경 개선･병력감축 대비한다" (2024)[URL]
- Akhtar, N., "Defense against universal adversarial perturbations", Proceedings of the IEEE conference on computer vision and pattern recognition, pp.3389~3398 (2018)
- Ali, M., "Adversarial attacks on ai based intrusion detection system for heterogeneous wireless communications networks", 2020 AIAA, 2020, 1~6
- Ban, Y., "Pre-trained adversarial perturbations", Advances in Neural Information Processing Systems, 35, pp.1196~1209 (2022)
- Carlini, N., "Extracting training data from large language models", 30th USENIX Security Symposium(USENIX Security 21), pp.2633~2650 (2021)
- Chakraborty, A., "A survey on adversarial attacks and defences", CAAI Transactions on Intelligence Technology, 6(1), pp.25~45 (2021)
- Chen, J., "De-pois : An attack-agnostic defense against data poisoning attacks", IEEE Transactions on Information Forensics and Security, 16, pp.3412~3425 (2021)
- DON Guidance on the Use of Generative Artificial Intelligence and Large Language Models, "Department Of Navy Chief Information Officer", The United State of America Departement Of Navy, ~ (2023)
- "Air Force releases new tool to track development, spending on AI efforts" (2024)[URL]
- Information Technology -Artificial Intelligence- Overview of Trustworthiness in Artificial Intelligence, "ISO", ISO, 2020
- "ISO/IEC JTC 1/SC 42 Artificial Intelligence" (2024)[URL]
- Lee, Y. B., "Test and Evaluation Procedures of Defense AI System linked to the ROK Defense Acquisition System", Journal of Society of Korea Industrial and Systems Engineering, 46(4), pp.229~237 (2023)
- Lewis, P., "Retrieval-augmented generation for knowledge-intensive nlp tasks", Advances in Neural Information Processing Systems, 33, pp.9459~9474 (2020)
- Long, T., "A survey on adversarial attacks in computer vision : Taxonomy, visualization and future directions", Computers &Security, 121, pp.102847~ (2022)
- National Cyber Security Centre, "Guidelines for secure AI system development" (2023), National Cyber Security Centre
- Ouyang, L., "Training language models to follow instructions with human feedback", Advances in neural information processing systems, 35, pp.27730~27744 (2022)
- Peri, N., "Deep k-nn defense against clean-label data poisoning attacks", Computer Vision–ECCV 2020 Workshops: Glasgow, pp.55~70 (2020), Springer International Publishing
- Radford, A., "Improving language understanding with unsupervised learning" (2018), OpenAI
- Tian, Z., "A comprehensive survey on poisoning attacks and countermeasures in machine learning", ACM Computing Surveys, 55(8), pp.1~35 (2022)
- Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence, "The Biden-Harris Administration", The United State of America: The Biden-Harris administration executive order, ~ (2023)
- "Guaranteeing AI Robustness Against Deception (GARD)" (2021)[URL]
- Wei, A., "Jailbroken: How does llm safety training fail?", Advances in Neural Information Processing Systems, 36 (2024)
- Zhou, M., "Dast: Data-free substitute training for adversarial attacks", Proceedings of the IEEE, 2020, 234~243