Against the backdrop of the deep integration of generative AI into music creation practices, this paper focuses on the modernization of traditional music in South Korea and China, a practice characterized by both high cultural semantic complexity and ...
Against the backdrop of the deep integration of generative AI into music creation practices, this paper focuses on the modernization of traditional music in South Korea and China, a practice characterized by both high cultural semantic complexity and cross cultural features. It aims to address the core question of whether the relationship between AI music generation and human music creation represents a difference in degree or a fundamental divide, and further explains why the phenomenon of “capability reversal” emerges in collaboration between humans and AI, as well as under what conditions such collaboration can serve as an effective pathway for reconfiguring the boundaries of creative capability. Centered on this core concern, the study analyzes four dimensions: how creators perceive and interpret “capability reversal”; how the boundaries of capability are constructed and evaluated within the distinct cultural contexts of South Korea and China; the mechanisms underlying the formation of “capability reversal”; and how collaboration between humans and AI addresses this phenomenon.
Adopting an interpretivist stance combined with a constructivist grounded theory approach, this study employs public text analysis, semi structured in depth interviews, and analysis of relevant literature on AI music generation technologies as its primary methods. Through continuous comparison and iterative coding of public narratives produced by South Korean and Chinese music creators, interview materials from Chinese music creators, and data on AI music generation technologies, the study establishes a chain of interpretation linking phenomena, contexts, mechanisms, and practical responses. Grounded in phenomenological intentionality theory, postphenomenological theory of technological mediation, cognitive musicology’s theory of musical structuration, and ethnomusicology’s theory of instrumental cultural codes, this study builds upon the SCReP theoretical framework presented in Chapter 2 to propose an empirical explanation of “capability boundary reconfiguration” at the level of creative practice.
The study finds that the intervention of AI music generation in the modernization of traditional music composition does not simply lead to an overall enhancement or replacement of creative capabilities, but rather facilitates a redistribution of different capabilities within the creative process across the stages of generation, screening, judgment, interpretation, and accountability. “Capability reversal” does not imply that AI is universally superior to humans in an abstract sense. Rather, under specific task conditions, AI demonstrates advantages in front end processes that are representable, repeatable, condition responsive, and open to rapid trial and error, while human creators remain irreplaceable in processes involving a high level of cultural semantic density, such as cultural interpretation, emotional meaning making, stylistic mastery, macrostructural coordination, aesthetic judgment, and accountability. A comparison between South Korea and China further demonstrates that the boundaries of capability are not fixed but are jointly shaped by factors such as cultural semantic complexity, task types, modes of media conversion, organizational structures, methods of technological intervention, and the allocation of cognitive resources among agents. On this basis, the practical potential of collaboration between humans and AI lies not in having AI replace humans in completing entire works, but in forming a collaborative structure characterized by layered cooperation, process based negotiation, and asymmetric responsibility. In this structure, AI is entrusted with draft generation, material exploration, and localized assistance, while human creators retain core competencies in meaning judgment, cultural translation, and final oversight.
The theoretical value of this paper is primarily reflected in three aspects. First, it proposes and elaborates the core concept of “capability reversal,” advancing research on creative relationships between humans and AI from static comparison toward a dynamic and contextualized analysis of capability boundaries. Second, building upon the SCReP framework, it introduces an explanation of “capability boundary reconfiguration” at the level of creative practice, revealing how structural conditions are transformed into mechanisms of capability allocation, responsibility attribution, and meaning negotiation within specific creative contexts. Third, it advances research on AI music generation from discussions centered on models and oriented toward outcomes toward interpretive pathways centered on practice, oriented toward process, and grounded in cross cultural analysis. The study also demonstrates that, within the context of the modernization of traditional music, the expansion of technologically generated capabilities has not eroded human agency. Rather, it has brought creative agency, aesthetic judgment, and cultural responsibility into sharper focus as nontransferable core competencies. Consequently, this paper not only provides a new analytical framework for understanding the reconfiguration of human and AI relationships triggered by the integration of AI music generation into traditional music creation, but also offers empirical grounding for the theoretical deepening, methodological refinement, and enhanced cultural sensitivity of research on AI music generation.