Eunsun Cho, Jong-Youn Rha
AI-driven recommendations help consumers efficiently find desired products, yet they can also act as external influences that intervene in consumer decision making. In the era of AI, it is essential for consumers to have opportunities to reflect their own values and desires, which is crucial for consumer-oriented AI technologies. This study examines the mediating role of consumer autonomy in the relationship between consumers' perceptions of AI-driven recommendation systems and their satisfaction. We propose that positive and negative perceptions influence satisfaction through consumer autonomy. In-depth interviews revealed ambivalent views regarding profiling for data input, algorithm as throughput, and recommended alternatives as output. Structural equation modeling showed that “alternative discovery,” “personalization,” and “manipulative practices” partially mediated consumer autonomy and satisfaction, while “information bias” and “privacy intrusiveness” were fully mediated by consumer autonomy. Consumer autonomy emerged as the most significant predictor of satisfaction, underscoring its critical role in shaping AI shopping experiences.