유럽평의회의 ‘인권·민주주의·법치주의 관점에서 본 위험 및 영향평가(HUDERIA)’가 「EU 인공지능법」상 ‘기본권 영향평가(FRIA)’에 미치는 영향 및 국내 시사점

Since the emergence of generative AI such as ChatGPT, the world has been enthusiastic about the positive opportunities that AI (hereinafter referred to as AI) will bring. However, the rapid development of AI has also presented challenges that humanity had not anticipated. Therefore, it is necessary to approach AI technology regulation in a different way than the traditional “permit first, regulate later” approach that has been applied in the development of existing technologies. The European Union (EU) has become the first in the world to enact comprehensive AI legislation, the EU AI Act, with the aim of preventing the dominance of big tech companies in the AI industry, fostering a fair competitive AI ecosystem, and preventing harm to human health, safety, and fundamental rights caused by AI. Under the EU AI Act, AI systems are classified into four categories based on their potential risk level, and high-risk AI systems are subject to a Fundamental Rights Impact Assessment (FRIA) under Article 27 of the Act. South Korea's AI Basic Act, enacted on January 21, 2025, and scheduled to take effect on January 22, 2026, also mandates a fundamental rights impact assessment for high-impact AI. However, it is currently unclear what specific provisions related to the fundamental rights impact assessment will be included in the implementing regulations of the AI Basic Act, which are expected to be announced by the Ministry of Science and ICT in June 2025. However, we can refer to specific details regarding fundamental rights impact assessments in European cases. This study aims to answer questions such as, “What is the background behind the introduction of fundamental rights impact assessments in the EU AI Act?”, “How did EU legislators structure fundamental rights impact assessments?”, and “What methodological standards should be followed when developing a fundamental rights impact assessment system?” The EU AI Act stipulates that the EU AI Office should provide more specific formats and standards for fundamental rights impact assessments, but the EU AI Office has not yet made any announcements on this matter. However, the European Commission, which must establish relevant implementing regulations, as well as companies, are already referring to “The Risk and Impact Assessment of AI Systems from the Point of View of Human Rights, Democracy and the Rule of Law(HUDERIA)” system proposed by the Council of Europe for AI to develop their own AI ethics guidelines. The implementation of EU regulations on high-risk AI systems is scheduled for August 2026, meaning that South Korea will implement the fundamental rights impact assessment mandated by law approximately six months before the EU. As such, South Korea could consider referencing the HUDERIA framework proposed by the Council of Europe, which the EU intends to adopt, to further concretize its own fundamental rights impact assessment system for high-impact AI.