Trust, Intention, and Behavioral Adaptation to Real Time Traffic Information Systems: A Latent Variable Model of Urban Mode Choice in Toronto

Abstract

The rapid diffusion of artificial intelligence (AI)-enabled real time traffic information systems is reshaping how urban travelers form perceptions, develop trust, and adapt their mode choice behavior. While existing transportation research has extensively examined the operational and network level impacts of intelligent mobility technologies, comparatively limited attention has been devoted to the latent cognitive and motivational mechanisms through which these systems influence individual travel decisions. This study develops and empirically tests an integrated behavioral framework that examines how AI information quality, system reliability, perceived safety, and accessibility perception are associated with system trust, travel intention, and self reported mode choice adaptation. Using survey data from 412 commuters in the Greater Toronto Area, Canada, the proposed model is estimated through Partial Least Squares Structural Equation Modeling (PLS-SEM) with bootstrapping and predictive validation procedures. The results indicate that system trust serves as a central mediating construct linking AI system attributes to travel intention, which in turn demonstrates a strong and systematic association with behavioral mode adaptation. The model exhibits substantial explanatory and predictive performance, with coefficients of determination reaching up to R² = 0.64 for key endogenous constructs. From a policy perspective, the findings highlight the role of trust centered and accessibility-oriented design of digital mobility platforms as mechanisms for reinforcing public transport strategies and promoting more sustainable travel behavior in metropolitan transport systems.