Interaction Design for Human-AI Choreography Co-creation

TL;DR

The paper addresses the problem of designing effective human-AI co-creation for choreography, an embodied domain requiring tailored interaction design. It synthesizes existing AI-assisted choreography work through three lenses—generation, creativity support, and co-creation—grounded in a COFI-based interaction-design framework to derive three key considerations: parallel collaboration, complementary human-AI roles, and rich communicative exchange. It maps an interaction design space across preparation and studio stages, detailing how collaboration and communication styles differ between ideation and prototyping, and proposes future directions for making AI a proactive partner. The contributions offer concrete design insights to advance co-creative choreography systems with real-time collaboration, embodied communication, and mutually intelligible human-AI dialogue, enabling more fluid and innovative performances.

Abstract

Human-AI co-creation aims to combine human and AI strengths for artistic results exceeding individual capabilities. Frameworks exist for painting, music, and poetry, but choreography's embodied nature demands a dedicated approach. This paper explores AI-assisted choreography techniques (e.g., generative ideation, embodied improvisation) and analyzes interaction design -- how humans and AI collaborate and communicate -- to inform the design considerations of future human-AI choreography co-creation systems.

Figures (1)

• Figure 1: Interaction design space for AI-supported choreography creation. The three axes are built upon the co-creative framework for interaction design introduced in rezwana2023designing.