개인정보 생애주기를 고려한 인공지능시스템 프라이버시 노출 대응 프레임워크
Privacy Exposure Corresponding Framework of Artificial Intelligence System Considering the Life Cycle of Personal Information
0
조회
0
다운로드
2
인용
0
참고문헌
키워드
초록
최근 챗봇을 통한 개인 프라이버시 노출 사고를 비롯하여 인공지능 시스템 활용단계에서의 개인 정보 노출 사고가 빈번해지고 있다. 유럽연합에서는 GDPR과 같은 제도적 장치를 통해 인공지능시스템에서 발생할 수 있는 개인 정보 노출을 제한하고 있지만, 실질적인 기술적 가이드는 미비한 상태이다. 본 연구에서는 인공지능 학습 및 활용 전주기 상의 프라이버시 침해 위험을 예방할 수 있는 프레임워크를 제안한다. 제안하는 프레임워크는 인공지능 시스템에서 개인 정보 수집부터 활용 및 폐기까지의 생애 주기 단계별로 적용 가능한 검증 기술로 구성되어 있다. 마지막으로 각 단계에서의 검증 기술이 적용된 사례를 살펴본다.
Abstract
Recently, privacy exposure at the stage of using artificial intelligence systems, including privacy exposure through chatbots, are becoming more frequent. Although the European Union limits the exposure of personal information that can occur in artificial intelligence systems through the institutional regulation called GDPR, practical technical guidance is insufficient. In this study, we propose a framework that can prevent the risk of privacy infringement throughout the lifecycle of AI learning and application. The proposed framework consists of verification technology that can be applied to each stage of the life cycle from the collection of personal information to the use and disposal of the artificial intelligence system. Finally, we look at the cases in which the verification technology is applied at each stage.
참고문헌 (25)
- "Boannews"[URL]
- "Gartner"[URL]
- Relevant ministries cooperation, "Artificial Intelligence (AI) Ethical Standards", ~ (2020)
- "Boannews"[URL]
- "iapp"[URL]
- "Hankyoreh"[URL]
- P. Romansky, "Challenges of the digitial age for privacy and personal data protection", Mathematical Biosciences and Engineering (2020)
- K. Benke, "Deep Learning Algorithms and the Protection of Data Privacy", JAMA Ophthalmol, pp.1024~1025 (2020)
- W. Yu, "Federated queries of clinical data repositories : balancing accuracy and privacy", Journal of Medical Internet Research (2020)
- F. Loukil, "Privacy-Preserving IoT Data Aggregation Based on Blockchain and Homomorphic Engryption", MDPI Sensors (2021)
- R. Shao, "Stochastic Channel-Based Federated Learning With Neural Network Pruning for Medical Data Privacy Preservation:Model Development and Experimental Validation", Stochastic Channel-Based Federated Learning With Neural Network Pruning for Medical Data Privacy Preservation: Model Development and Experimental Validation, ~ (2020)
- Soo-Yong Shin, "Privacy Protection and Data Utilization", Privacy Protection and Data Utilization, ~ (2021)
- H. M. Verhelst, "Machine Learning Against Terrorism: How Big Data Collection and Analysis Influences the Privacy-Security Dilemma", Science and Engineering Ethics (2020)
- T. Wang, "A Comprehensive Survey on Local Differential Privacy toward Data Statistics and Analysis", MDPI Sensors (2020)
- L. Yao, "Sensitive attribute privacy preservation of trajectory data publishing based on I-diversity", Distributed and Parallel Databases (2020)
- M. Asad, "A Critical Evaluation of Privacy and Security Threats in Federated Learning", MDPI Sensors (2020)
- M. Nasr, "Comprehensive Privacy Analysis of Deep Learning : Passive and Active White-box Inference Attacks against Centralized and Federated Learning", 2019 IEEE Symposium on Security and Privacy (2020)
- 신영진, "우리나라의 인공지능(AI)서비스를 위한 개인정보보호 개선방안", 융합정보논문지, 11(3), pp.20~33 (2021), 중소기업융합학회[DOI]
- Jong-Seon Kim, "Differential Privacy – Privacy Protection Big Data Processing Standards", The Journal of The Korean Institute of Communication Sciences, pp.18~24 (2020)
- 홍은주, "신경망 학습에서 프라이버시 이슈 및 대응방법 분석", 디지털융복합연구, 17(7), pp.285~292 (2019), 한국디지털정책학회[DOI]
- Gwon-Sang Ryu, "Artificial intelligence security attacks and countermeasures research trends", Journal of Korea Institute of Information Security & Cryptology (2020)
- Samsung SDS, "Samsung SDS pursues Privacy Enhancement Technology", ~ (2021)
- "Google"[URL]
- R. Shokri, "Membership inference attacks against machine learning models", 2017 IEEE Symposium on Security and Privacy (2016)
- Ministry of the Interior(MOI), "Personal information protection guidelines for system development and operator", ~ (2015)