Dark Side of AI in online shopping: does algorithmic targeting increase unsustainable consumption

Abstract

This study investigates the dark side of AI-targeted online shopping by examining how algorithmically curated stimuli influence consumers’ internal states, impulse buying urge, and unsustainable consumption. Guided by the Stimulus–Organism–Response (S-O-R) framework, streamer interaction, peer interaction, realness, and vividness are conceptualized as environmental cues activating perceived usefulness, trust in recommendation, arousal, perceived ease of use, perceived uncertainty, and perceived manipulation. Using a sample of 642 Vietnamese online shoppers and PLS-SEM analysis, results indicate that impulse buying urge is significantly driven by perceived usefulness, trust in recommendation, and arousal, whereas perceived ease of use, uncertainty, and manipulation show non-significant effects. Impulse buying urge, in turn, positively contributes to unsustainable consumption, highlighting a behavioral pathway through which AI-enabled shopping may promote low-necessity purchases. These findings extend the S-O-R model to negative outcomes of AI-mediated retail and clarify the roles of recommendation trust and affective activation. Managerial implications suggest that online retailers and AI platform designers implement responsible recommendation practices, transparent personalization cues, and reflective purchasing interventions to mitigate unsustainable consumption in digital commerce.