Deep versus superficial anthropomorphism: Exploring their effects on human trust in shared autonomous vehicles

Recently, there is an increasing trend to study the impact of anthropomorphism on human trust. However, previous research has shown divergent results. In this study, shared autonomous vehicles (SAVs), an emerging mobility solution, are used to explore the threshold effects of anthropomorphism on human trust. Structural equation modelling is deployed to test the research model based on a valid survey sample of 451 respondents. Results show that a superficial level of anthropomorphism can significantly boost human trust via interaction quality and facilitating condition. By contrast, a deep level of anthropomorphism would decrease human trust. Although the negative effect does not reach statistical significance based on the whole sample, results of the multigroup analysis show that a deep level of anthropomorphism has a significant negative effect on human-SAV interaction quality when respondents possess the following characteristics: (1) male, (2) low-income, (3) low-education, or (4) no-vehicle ownership. Regarding theoretical contribution, this study enriches the literature by identifying the threshold effects of anthropomorphism on human trust. Regarding policy and management implications, this study offers some implications on adding anthropomorphic features to SAVs.