Y. Cao, Fang Wang
Generative AI technologies are reshaping online virtual personas with hyper-realism, accelerating their adoption across online business contexts. This research investigates audience responses to hyper-realistic generative AI streamers versus human streamers within livestream commerce, drawing on source credibility theory to explain potential differences. Across four online scenario-based experiments, we find that audiences generally exhibit lower purchase intentions toward AI streamers compared to their human counterparts. This disparity is largely attributed to the lower perceived credibility of AI streamers. Moreover, this credibility gap is moderated by audience traits, such as novelty-seeking and social overload, with AI streamers appealing more to audiences characterized by higher levels of these traits. This research enriches scholarly understanding of audience response to AI-generated personas in dynamic online environments and offers practical insights for marketers deploying AI streamers in their businesses.