An Interaction Between Anthropomorphism and Personality on Trust in Automated Systems

Automated assistance is increasingly being implemented in domains ranging from healthcare to transportation. The reason for the tendency for certain users to trust or mistrust automated assistance has been studied to mixed effect. I examined the effect of anthropomorphism as an independent factor on user trust. In addition, I examined the potential for anthropomorphism to act as a moderator between the personality traits of a user and the trust a user demonstrates in the automated aid. Though the participants in the anthropomorphic condition did view the assistant as more human-like, the level of anthropomorphism had no effect on user behavior. The traits that have been previously found to have an effect on interaction with an automated assistant had their impacts reversed. Users high in extraversion and trait trust were less likely to display trusting behaviors when dealing with an anthropomorphized automated assistant. This expands trait activation theory to the domain of automated interaction. It also allows for a more nuanced understanding of user-automation interaction that impacts selection for any position that interacts with an automated assistant