인공지능이 인권과 민주주의에 미치는 영향과 인공지능법안의 쟁점

Abstract

Artificial intelligence (AI) encompasses both positive and negative aspects from the perspectives of human rights and democracy. AI not only has the potential to radically transform political, economic, and social structures but also raises new constitutional and legal issues. This article focuses on the political and legal awareness and examines the impact of AI on human rights and democracy, as well as highlights the key points of contention in recent discussions of AI legislation in the legislature. AI systems can have detrimental effects on nearly all human rights and fundamental freedoms. Particularly, they can significantly impede democratic processes by undermining freedom of expression and the right to self-determination of personal information. “Content moderation” and “user profiling” serve as representative examples that illustrate the risks posed by AI systems to fundamental rights and democracy. In the same vein, the regulation and management of misinformation online become contentious issues. Uncontrolled AI systems not only render individuals vulnerable to human rights violations but also jeopardize the fate of political communities by placing decision-making power in the hands of private enterprises. Moreover, as surveillance societies become more sophisticated and differentiated, traditional surveillance powers evolve into big data surveillance powers, where surveillance authority solidifies as a social infrastructure rather than an antagonistic object of human rights subjects. Under surveillance capitalism, individuals are depersonalized, fragmented, and reduced to data fragments. Freedom is captured through delegated decision-making based on “consent” and monopolistic control of information sources by AI systems, while individual and group identities are manipulated algorithmically, leading to polarization. Freedom is not externally constrained but internally distorted. The AI society enables private actors such as big tech companies and digital platforms to exercise private power beyond the role of the state or government, challenging constitutional traditions. Many countries around the world are preparing legislation to address the potential risks of AI technologies. National Assembly of South Korea is also undergoing the legislative process regarding AI bills. However, the “permit-first-regulate-later principle” adopted by these bills is inadequate to address the inherent risks associated with AI technologies and is not necessarily favorable to industrial growth. Furthermore, unlike the detailed classification systems in the European Union and elsewhere, these bills fail to regulate prohibited AI and unreasonably exclude areas with significant risks to life, safety, and fundamental rights without sufficient justification. Additionally, by designating the Ministry of Science and ICT, a technocracy-centered institution, as the main governing body for AI policies, the current bills emphasize an industry-biased AI policy governance. Therefore, the pending bills require comprehensive reconsideration.