Abstract:

This paper uses practice theory to reveal the role of movement in the qualification of a digital-human duet. We examine the co-design of a generative AI dance partner to explore how movement becomes a practical site where value, responsibility, and what it means to partner are worked out. Drawing on ten months of iterative development and testing, centred on nine extended studio-based co-design sessions, we trace how the AI-partner is trained, rendered, and reworked in practice. We show how the partner is made move-with-able through three linked episodes: holding contradictory evaluations (creative promise alongside technical failure), making the relation readable enough to respond in time (through rules, simplified forms, and sensory cues), and shifting toward movement co-proposed (initiative, interruption, humour). At the same time, XR conditions and model sensitivities narrow the human dancer’s repertoire. The findings advance work on qualification by introducing kinaesthetic qualification – sensorimotor practice as a way of working out what matters, how, and to whom – and show how distributed care work, spanning data stewardship, calibration, rendering, and boundary-setting, makes moving-together workable over time.

Please choose your submission type: Full Paper

Primary Track: Practice Theory

Menu

Ms. Yinan Liu