인공지능 규제

Abstract

Artificial intelligence (AI), the key technology in the Fourth Industrial Revolution, is expected to bring numerous socioeconomic benefits, but there are also concerns about its adverse effects. This study aims to explore policy options for building trust in AI and discuss their limitations. Specifically, we analyze the EU and US approaches to creating so-called “trustworthy AI” and draw policy implications for South Korea. The EU's approach to overcoming differences in socioeconomic maturity, government capacity, and institutional context among its member states is both normative and centralized. On the other hand, the United States, which aims to maintain its technological hegemony, has taken a light-touch approach to promoting the AI industry by granting discretionary authority to sectoral regulators while encouraging private sector autonomy. The changing environment calls for a full-scale discussion on AI regulation in Korea. In this paper, we point out that it is essential to consider regulatory arrangements that include at least three elements: risk-based regulation, co-regulatory framework, and harmonization with international standards. We conclude by discussing policy options and limitations.