Heterogenous evaluations of autonomous vehicle services: An extended theoretical framework and empirical evidence

Abstract

Amidst rising interest in autonomous vehicle services, extant literature reveals a paucity of research examining: 1) both cognitive and emotional evaluations; 2) characteristics of the service context (e.g. risk) and artificial intelligence (e.g. autonomy); and 3) heterogenous outcomes. Moreover, there are mixed findings on autonomous vehicle adoption/resistance. To address these gaps, we develop the Customer Responses to Unmanned Intelligent-transport Services based on Emotions and Cognitions (CRUISE-C) framework by extending the earlier CRUISE framework and building on the Elaboration Likelihood Model. The framework further delineates four segments, which differ in cognitive and emotional evaluations of fully autonomous vehicle services. We test CRUISE-C using three experimental studies. Study 1 shows that the resistant segments consider fully autonomous (vs. regular) vehicle services to be more vulnerable, and less reliable and convenient. Study 2 shows that a service failure involving fully autonomous (vs. regular) vehicles does not increase negative emotions in any segment, but attenuates perceived severity in Segment 1 (“Avoiders”) and slightly amplifies perceived severity in Segment 4 (“Aficionados”). In Study 3, a machine learning model reveals segment membership as the strongest predictor of individuals' readiness to adopt autonomous vehicles, followed by reliance on taxis, female vs. male, and cognitive and emotional evaluations.