시몽동의 구체화 이론을 통해 본 인공지능 생성예술: 구체화의 적용 가능성과 한계 탐색 = AI Generative Art through Simondon's Concretization Theory: Exploring the Applicability and Limitations of Concretization

다국어 초록 (Multilingual Abstract)

This study began by analyzing the development of AI technology and the creative process of AI-generated art through the framework of Gilbert Simondon's philosophy of technology. The purpose of this study was to critically explore the possibilities and limitations of applying Simondon's theory of concretization to the development of AI technology and AI generative art. Therefore, this paper shifted to examining how the process of concretization, in which a technological object evolves from an "abstract mode" to a "concrete mode," as Simondon described it, fits with AI technology and AI-generated art. The results and contents of this study are as follows. First, while AI technology has advanced by leveraging computing resources and big data, it differs from Simondon's concept of concretization in that the learning and inference stages are clearly separated and the inference stage operates as a fixed mathematical function. Second, the prompt-based conditioning in AI generative art fundamentally differs from Simondon's concept of associated millieu. Third, collaboration between artists and AI should be reconceptualized as a human-AI collaborative system, rather than the formation of associated millieu. Based on the results of this study, we demonstrate that AI is a new form of technological entity distinct from both physical technological objects and traditional tools. We also hope to critically apply Simondon's theory of concretization to AI generative art.