Development and Validation of an AI Virtual Streamer Scale for Live-Streaming E-Commerce

In recent years, AI virtual streamers (i.e., digital characters with human-like appearances) have become a pervasive e-commerce phenomenon but remain almost unexplored in the literature on e-commerce marketing. The possible reasons for this deficiency are the vagueness of the AI virtual streamer concept and the lack of a measurement tool. Therefore, this study aimed to conceptualize AI virtual streamers and develop an AI virtual streamer characteristic scale through five stages based on social response theory. Following item generation, measurement refinement, exploratory factor analysis, confirmatory factory analysis and nomological validity, a third-dimensional (persona, anthropomorphism, and interactivity) scale with 10 items was established. Empirical results reveal sound measurement qualities in terms of reliability and validity, and demonstrate its unique position in respect to established marketing concepts. The developed scale has managerial implications for designing and applying virtual humans or avatars, such as virtual employees, virtual spokespersons, virtual idols, and virtual guided tours. This tool also enables future research on virtual live streaming, such as virtual destination live streaming, virtual education live streaming, or virtual live concerts.