Practice and Reflection on AI Motion Capture Technology in Dance Creation

Abstract

With the rapid development of artificial intelligence (AI) technology, its involvement in the field of dance creation has become increasingly profound. In particular, within dance education and teaching, AI not only serves as a tool-based assistant but is also gradually demonstrating its potential as a co-creative subject. Taking AI motion capture technology as the point of entry, this study explores the evolving role of AI in contemporary dance creation and offers critical reflections on its implications. Three dance performance videos with fast, medium, and slow paces were selected, and corresponding 3D motion data were generated via the DeepMotion platform to conduct experimental creative practices. By comparatively analyzing the original dance performances and the AI-generated outputs, this study reveals the characteristics of algorithmic performance in motion recognition, rhythmic logic, and stylistic construction, while reflecting on how “non-human choreographic thinking” expands the possibilities of dance reconstruction and the boundaries of creation. The research aims to advance dance studies within the context of human–AI co-creation, considering how dance in the technological era can retain its corporeality and cultural essence while moving toward a more open and cross-disciplinary future.