인공지능 기술 규제에 관한 탐색적 고찰: 딥러닝 기술을 중심으로

In this study, we explore how to regulate Artificial Intelligence (AI). Though the interest in AI has surged recently, an agreement on how to regulate AI has not yet reached due to the different understanding on AI. We first thoroughly study prior arts and national policies, and analyse technologies, including deep learning in detail. Then, we set the goal of AI regulation as securing transparency, fairness, safety, and accountability. The scope of regulation is set up to the AI system, the development process and utilization. We consider developers and users as the subject to regulation. Then, we introduce practical AI regulation methods and make some proposals for regulation authorities. An academic significance of this study is to lay the solid foundation for the consistent discussion on AI regulations by analyzing the current technical level of AI. However, it needs to be mentioned that the verification of this study should be followed in the future by actually applying it to the on-site regulations for each field of AI, since this study explores the direction of AI regulation from a general perspective.