한· 싱가포르 규제샌드박스 비교를 통한 AI 샌드박스 효율화 방안 제언

Regulatory sandboxes have served as institutional tools that allow new products and services based on emerging technologies to be introduced into the market under limited conditions, granting temporary exemptions from existing regulations. These trials are used to test and validate innovations, and the results are reflected as a basis for regulatory improvement and broader market adoption. In recent years, as artificial intelligence (AI) technologies have increasingly converged with various industries, AI-related businesses have been designated across multiple regulatory sandboxes in Korea, including those for industrial convergence, ICT, smart cities, and innovative finance. However, the current system faces structural challenges. Each sandbox is managed by a different government ministry based on the industry sector, which can result in a mismatch between the technical characteristics of the proposed innovation and the regulatory expertise of the overseeing authority. This misalignment can lead to delays in regulatory improvements or impose excessive compliance conditions that undermine the commercial viability of new technologies. To explore alternatives to these limitations, this study examines Singapore’s AI-specific regulatory sandbox model, which is led by the Infocomm Media Development Authority (IMDA). IMDA operates a sandbox to assess the safety and trustworthiness of generative AI systems, aiming to establish a structured framework for pre-deployment testing and evaluation. At the same time, it also runs sandboxes designed to encourage AI adoption among small and medium-sized enterprises (SMEs). This dual structure—combining government-led regulatory design with participation from the private sector—is regarded as an effective strategy to bridge emerging technologies with responsive regulatory frameworks. The paper argues that Singapore’s model offers valuable insights for Korea, especially in light of the recently enacted Basic Law on AI. To move beyond fragmented, sector-specific regulation, Korea should establish a dedicated sandbox framework for general-purpose technologies like AI, incorporate top-down regulatory planning, and build an integrated management system under a centralized regulatory authority. By grounding regulatory design in empirical evidence from pilot testing, Korea’s regulatory sandbox system can evolve into a dynamic and strategic platform capable of addressing the wide-ranging risks and impacts of advanced AI technologies.