Taking the Tujia hand-waving dance as an example, this study aims to apply AI technology to design to improve the design efficiency of intangible cultural heritage and promote the development and communication of related digital collections. Firstly, based on the video of inheritors’ performance, the study establishes a semantic dataset of the typical postures of Tujia hand-waving dance. Then, the study tries 3D reconstruction of the human body from the single-view images using AI algorithms. Later, the designer carves the details of the 3D models generated by AI. Finally, the human-AI co-created 3D models are transformed into NFT works. This study initially confirms that not only the AI-assisted workflow may obviously reduce the workload that designers need to invest in the early stage of modeling, but also can efficiently connect the links of data collection, 3D modeling, detail carving, and morphological transformation, thus improving the design efficiency of intangible cultural heritage.