Impromptu Drama in Human-Machine Collaboration: The Innovative Practice of Artificial Intelligence in Chinese Theatre
List of Authors
Jingying Zhang, Lanbin Liu, Syahrul Fithri Musa
Keyword
Improvisational Theatre; Artificial Intelligence; Chinese Theatre; Creativity; Audience Engagement
Abstract
Improvisational theatre, characterised by spontaneous interaction between performers and audiences, often faces creative fatigue, formulaic output, and limited feedback. This study explores how generative AI might alleviate these issues and inform Chinese improvisational practice by analysing the Improbotics system—whose prompts evolved through successive iterations (A.L.Ex → GPT 2 → GPT 4 → Llama v2). Drawing on publicly available literature and video recordings of live demonstrations and workshop sessions, thematic analysis was applied to 257 AI cue events, triangulating document insights with video observations. The findings demonstrate that prompts become steadily more novel and contextually coherent with each model upgrade, markedly reducing performers’ cognitive load, enhancing their adaptive responses, and deepening audience engagement as evidenced by laughter, applause, and reflective silence in the recordings. The study also identifies simple compensatory mechanisms (peer cues, hand signals) used to preserve performance flow when prompts are ambiguous. These results suggest that integrating advanced AI prompts into improvisational theatre can sustainably broaden creative horizons, refine rehearsal and performance methods, and support wider artistic and commercial innovation.