Uncanny Valley Effects on Chatbot Trust, Purchase Intention, and Adoption Intention in the Context of E-Commerce: The Moderating Role of Avatar Familiarity

Abstract

This study investigates the effect of chatbot humanization on the perception of eeriness, trust, and users’ behavioral intention. Specifically, this study employed a 2 (humanization of chatbot agent avatar: hyperrealistic-animated vs. cartoonish-still) × 2 (avatar familiarity: celebrity avatar vs. non-celebrity avatar) between-subjects experiment (N = 185), in which participants were asked to purchase a laptop from an e-commerce vendor by interacting with a chatbot agent. Based on predictions from the uncanny valley effect hypothesis (UVE), enhancing the human likeness of a chatbot agent through visual realism and animacy was predicted to negatively influence users’ trust in the chatbot agent and behavioral intention as a consequence of the activation of a negative affective state (i.e., a feeling of eeriness). Consistent with our predictions, the results from PLS-SEM showed that (a) enhancing the human likeness of a chatbot agent significantly increased users’ feeling of eeriness, (b) the feeling of eeriness negatively influenced users’ trust in the chatbot agent, c) trust, determined by the feeling of eeriness, significantly affected users’ purchase intention and willingness to reuse the chatbot, and d) the relationship between humanization and eeriness significantly moderated by the familiarity of the chatbot avatar. We discuss the theoretical implications of the current study on UVE as well as its practical implications for the implementation of anthropomorphized chatbot agents in the e-commerce context.