Regulatory Trends and Legal Challenges in the AI-based medical industry

This Thesis deals with the development of AI-based medical industries and the regulatory and legal challenges surrounding them. While AI has the potential to improve efficiency in healthcare by supporting diagnosis, data analysis, and treatment across various areas, ethical and legal concerns, such as privacy protection, bias, discrimination, and legal liability, have emerged. To address these, countries are moving forward with AI-related regulations, and this paper analyzes and compares the legislative developments in the United States, the EU, and South Korea. In the U.S., efforts to increase AI fairness and transparency include the Algorithmic Accountability Act and the adoption of innovative medical device programs to regulate AI medical devices. The EU, in 2024, enacted the AI Regulation (AI Act), which categorizes AI risks into four levels and imposes strict regulations on high-risk AI, including in the medical field. South Korea, in turn, has introduced guidelines for the approval and review of AI medical devices and is actively promoting legislation to ensure AI reliability and safety. Despite these regulatory advancements, unresolved issues remain. First, when patients are harmed due to the diagnostic results provided by AI medical devices, establishing clear legal responsibility is essential. Second, large volumes of personal data are used in AI training, leading to challenges related to the processing and reuse of pseudonymized data. Although a recent amendment to South Korea’s Personal Information Protection Act grants individuals rights over automated decisions, the specifics of explanation requirements remain unclear. Third, as AI learns from historical data, there is a risk of biased conclusions or discrimination against certain groups, underscoring the need for measures to ensure AI reliability and fairness. Finally, as AI advances to a stage where it can make independent decisions, there is a need to consider granting legal status to AI. The emergence of Artificial General Intelligence (AGI) may further necessitate a fundamental review of whether AI should hold legal personhood and if so, how to define the subject of medical acts and assign corresponding liability. This paper underscores the importance of balancing the growth of AI medical technology with appropriate regulatory safeguards and highlights the necessity for dedicated legal frameworks in healthcare to address these unique challenges.