An Empirical UX Study of an AI Control Model for Audience-Interactive Smartphone Performance Design

The widespread adoption of smartphones and wireless networks enables audiences to actively participate in performance production; however, conventional stage control systems remain operator-centered and struggle to stably reflect large-scale audience inputs. Network congestion and device heterogeneity lead to end-to-end latency and execution variance, degrading the quality of service (QoS) in audience-interactive performances. To address these issues, this study designs an AI-based adaptive control model that generates performance-appropriate control parameters from audience interaction data and validates it through empirical experiments conducted in a real performance venue. The results confirm that stable QoS can be maintained in multi-device environments using real-time communication and seat-based mapping, demonstrating the effectiveness of integrating AI-based control into the performance production process to enable audiences to participate as real-time co-creators.