To Be Human-like or Machine-like? An Empirical Research on User Trust in AI Applications in Service Industry

Abstract

This research investigates user trust in artificial intelligence (AI) applications based on empirical analysis of intelligent service robots in service industry. We examined the impacts of human-like design of AI robots on intelligence and anthropomorphism perceived by users. The relationships between perceived intelligence, perceived anthropomorphism, and level of trust were also tested. An online survey with a between-subject design was conducted to collect data. The results indicate that robots with physical human-like appearance were perceived lower level of anthropomorphism and intelligence, and interaction function design on robots does not significantly increase those perceptions. Moreover, the significant effect of perceived intelligence on perceived anthropomorphism was confirmed, and both perceptions of users exert positive effect on their trust in AI robots.