Human likeness and behavioral authenticity in virtual influencers: Effects on consumer responses and underlying mechanisms

With the growing popularity of computer-generated social media personas, known as virtual influencers, in brand communication, companies are gaining innovative marketing tools while simultaneously facing the challenge of understanding how consumers perceive these non-human endorsers. Drawing on the Uncanny Valley Theory, Parasocial Interaction Theory, and the Technology Acceptance Model, this study develops a dual-cue analytical framework to systematically examine the joint effects of virtual influencers’ human likeness and behavioral authenticity on brand attitude.Based on three pre-registered randomized experiments (N = 374, 316, and 494, respectively), the study yields the following findings. First, human likeness exhibits an inverted U-shaped relationship with brand attitude, with a mediate level of human resemblance generating the most favorable evaluations. Second, parasocial interaction and perceived credibility sequentially mediate the effect of human likeness on brand attitude, while behavioral authenticity mediates this mediation process. Under conditions of low behavioral authenticity, the mediation effect becomes fully dependent on these two mechanisms. Third, consumers with higher AI readiness show greater acceptance of highly anthropomorphic virtual influencers; however, their sensitivity to insufficient behavioral authenticity does not diminish. Instead, they become even more critical of authenticity deficiencies. Furthermore, the effects of virtual influencers on purchase intention and positive word-of-mouth intention are fully mediated by brand attitude.The study makes several theoretical contributions. It extends the cross-cultural applicability of the Uncanny Valley Theory, reveals the threshold effect of behavioral authenticity, and identifies the multidimensional structure of AI readiness. From a managerial perspective, the findings suggest that brands should adopt a “mediate anthropomorphism–high behavioral authenticity” strategy, avoid the “high anthropomorphism–low authenticity” combination, and develop differentiated marketing approaches for consumer segments with diverse digital-cultural backgrounds.