Consumer Trust and Privacy Concerns in AI-Driven E-Commerce: Evidence from a Hybrid SEM–ANN Study

Framing the study within the Unified Theory of Acceptance and Use of Technology (UTAUT) and the extended online privacy concern model, this study investigates the effects of effort expectancy, social influence, privacy concerns, and perceived risk on e-commerce consumer behavior, while positioning trust in AI as a key mediating factor. The data were collected from 250 active e-commerce users in Bangladesh, and the analysis was done with a hybrid approach that integrates partial least squares structural equation modeling (PLS-SEM) with artificial neural network (ANN) analysis. The SEM results indicate that effort expectancy and social influence have significant positive effects on consumer behavior. Social influence also shows a significant positive effect on trust in AI. Trust in AI exhibits a significant positive effect on consumer behavior. However, perceived risk does not show a significant effect on either trust in AI or consumer behavior. Privacy concerns demonstrate a significant positive relationship with both trust in AI and consumer behavior, contrary to the hypothesized negative relationships. The mediation analysis shows that trust in AI significantly mediates the relationship between social influence and consumer behavior, while no significant mediation effects are observed for privacy concerns or perceived risk. The ANN results further confirm the dominance of social influence as the most important predictor of both trust in AI and consumer behavior, followed by effort expectancy and trust in AI, while perceived risk shows minimal predictive relevance. Overall, the findings suggest that consumer adoption of AI-enabled e-commerce is primarily driven by benefit-oriented factors rather than risk-based considerations in the present context. The study contributes to the literature by extending UTAUT and privacy calculus theory to AI-mediated commerce and by demonstrating the value of combining SEM and ANN to capture both explanatory relationships and predictive importance. From a managerial perspective, the results highlight the importance of strengthening social influence mechanisms, improving system usability, and building trust in AI systems to enhance consumer engagement in AI-driven e-commerce environments.