A Constitutional Analysis of Fundamental Rights in Relation to High-Impact Artificial Intelligence - With a Comparative Study of the EU AI Act and the Framework Act on the Promotion of Artificial Intelligence and the Establishment of Trust in Korea

This paper examines the constitutional implications of the term “High-Impact Artificial Intelligence” (HI-AI) as defined in Korea’s Framework Act on the Promotion of AI and Establishment of Trust (2024). With the rapid development of generative AI and its increasing integration into high-risk sectors such as healthcare, public administration, finance, and transportation, concerns about potential infringements on fundamental rights—such as human dignity, privacy, and equality—are intensifying. In response, this study analyzes the legal responsibilities and regulatory framework imposed on HI-AI in Korea’s framework law, while engaging in a comparative legal analysis with the European Union’s Artificial Intelligence Act, the world’s first comprehensive AI regulatory regime. The EU AI Act adopts a risk-based classification model, introducing strict prohibitions and pre-market requirements for AI systems that pose unacceptable or high risks to fundamental rights. In contrast, Korea’s framework law employs the term “high-impact” rather than “high-risk,” emphasizing regulatory management through transparency, user protection, and human oversight without corresponding enforcement measures or sanctions of comparable magnitude. While the Korean law is more industry- friendly and oriented toward technological promotion, it faces criticism for its limited legal enforceability and potential gaps in rights protection. From a constitutional perspective, HI-AI systems can affect core rights such as the right to human dignity, the right to informational self-determination, the right to equality, the freedom to choose one’s occupation, and due process. Particularly problematic are issues surrounding algorithmic bias, accountability, lack of explainability, and limited mechanisms for legal redress. As such, this study argues that constitutional legitimacy should form the foundation of any regulatory framework for HI-AI and recommends enhancements to ensure transparency, enforceable rights to explanation and objection, and legal certainty in governance structures. Ultimately, the paper advocates for a balanced regulatory approach that fosters innovation while firmly anchoring AI development within a human-centered constitutional framework. It underscores the necessity for robust institutional safeguards that harmonize AI advancement with the protection of individual rights in the digital age.