Too Human to Trust? Exploring the Dark Side of Anthropomorphic Service Robots Through Identity Threat and Social Anxiety

With the increasing integration of artificial intelligence into service industries, understanding consumers’ psychological responses to anthropomorphic service robots has become increasingly important. This study aims to investigate how perceived anthropomorphism influences consumers’ usage intention by incorporating identity threat as a mediating mechanism and social anxiety as a boundary condition. Across three studies, we systematically tested the proposed conceptual model. Study 1 employed a survey-based approach and linear regression analysis to examine the direct effect of perceived anthropomorphism, revealing that higher levels of anthropomorphism significantly increased consumers’ willingness to use the service. Study 2 introduced identity threat as a mediator and confirmed its partial mediating role, suggesting that while anthropomorphic features may enhance user engagement, they can also evoke discomfort by threatening human distinctiveness, thereby reducing usage intention. Study 3 incorporated social anxiety into a moderated mediation model using PROCESS Model 7. Results showed that the indirect effect of anthropomorphism on usage intention through identity threat was significant only among participants with low social anxiety, whereas the effect was non-significant for those with high social anxiety. These findings contribute to the literature on human–AI interaction by revealing the double-edged effects of anthropomorphism and highlighting how individual psychological traits shape user acceptance of service robots.