Generative AI 기반 공간디자인 교육실행 사례분석 및 효과성 연구

Abstract

(Background and Purpose) Usually, the use of Generative AI (GenAI) in spatial design education has been restricted to introductory instruction on tool functions or stage-specific applicability. In particular, studies on integrating with the Double Diamond model have also focused on exploring the applicability, leaving few empirical studies on its application in actual educational field. Therefore, this empirical study differentiated itself from existing research by applying the educational process to actual university classes, designing and implementing a curriculum, and analyzing learner outcomes. The purpose of this study was to identify the effectiveness and limitations of GenAI-based education and to explore directions for the development of future educational systems.(Method) A hands-on curriculum was designed for the senior students majoring in design in the first semester of the 2025 academic year and was applied to their actual classes. The roles of text-based and image-based GenAI were differentiated and systematically integrated throughout the curriculum. Total five teams participated in the class, and the overall learning process and outcomes were systematically recorded and analyzed. Through this process, the step-by-step educational effects and the limitations and implications revealed during the class were derived. (Results) GenAI demonstrated positive effects in supporting learners throughout the entire curriculum, but it also revealed new limitations and areas for refinement at each stage. In the Discover phase, text-based GenAI facilitated access to information and encouraged divergent thinking, but the need for training in data reliability verification and critical selection emerged. In the Define phase, concepts were clearly established with text-based GenAI supporting strategic ideation and image-based GenAI supporting conceptual visualization; however, critical evaluation of AI-generated materials became necessary. In the Develop phase, the complementary use of both AI types enhanced creative exploration and critical selection, yet over-simplification of proposals required consistency checks with drawings and technical details. In the Deliver phase, text-based GenAI improved the completeness of presentation materials, but over-reliance on AI raised concerns about diminishing students’ independent judgment. (Conclusions) This study has shown that generative AI goes beyond simple instrumental assistance in spatial design education, showing significant positive effects in areas such as information search efficiency, concept visualization and creative thinking. However, limitations were also revealed, requiring improvement in areas such as information reliability verification, visualization-drawing connectivity, and maintaining student agency. Therefore, the significance of this study lies in specifying a practical operational model for GenAI-based spatial design education and providing basic data for future extensive verification in various educational contexts by comprehensively presenting these benefits and limitations.