As artificial intelligence technologies continue to be embedded in digital platforms, recommendation mechanisms in social commerce are evolving beyond traditional information matching toward intelligent recommendation mechanisms based on data identifi...
As artificial intelligence technologies continue to be embedded in digital platforms, recommendation mechanisms in social commerce are evolving beyond traditional information matching toward intelligent recommendation mechanisms based on data identification, preference prediction, and contextual fit. Compared with general information provision, AI recommendations not only lead consumers to subjectively perceive the relevance and fit of recommended content, but also exert a deeper influence on their emotional experiences and relational judgments toward the platform. However, AI recommendations do not always produce stable positive outcomes. On the one hand, when recommended content better matches consumers’ needs, consumers are more likely to form a subjective experience of being understood, responded to, and effectively served. On the other hand, as recommendations become more sophisticated, consumers may become more aware that the platform is continuously collecting, analyzing, and using their personal information. As a result, privacy concerns may be triggered, thereby weakening the positive effects of personalized experiences. Against this background, this study systematically examines the internal psychological mechanism through which AI recommendations influence consumer loyalty, as well as its boundary condition, in the context of Chinese social commerce.
This study adopts the Stimulus-Organism-Response theory as the overall analytical framework and integrates self-determination theory with discussions related to the personalization-privacy paradox. Based on this theoretical foundation, this study constructs a moderated serial mediation model of “perceived personalization–resonance–happiness–loyalty.” In this model, perceived personalization serves as the key psychological starting point through which the effects of AI recommendations operate. Resonance and happiness represent the sequential internal psychological states experienced by consumers in the platform recommendation context. Loyalty is reflected in behavioral tendencies such as continued use, preferential choice, positive recommendation, and relationship maintenance. Privacy concern is positioned as a key boundary condition that explains why consumers may respond differently under conditions where personalization benefits and privacy risks coexist.
In terms of research design, this study adopts a progressive experimental research approach and systematically verifies the theoretical model through one preliminary study and three main studies. The preliminary study was used only to verify the validity of the privacy concern manipulation materials used in Study 3. Study 1 examined the core psychological path and serial mediation mechanism in the context of vacation hotels, a high-involvement hedonic product. Study 2 further tested the robustness of the core mechanism across product types in the context of laptops, a high-involvement utilitarian product. Study 3 examined the boundary effect of privacy concern more rigorously by manipulating both perceived personalization and privacy concern in the context of skincare products, a low-involvement hedonic product. In all three main studies, manipulation checks were conducted to confirm the validity of the experimental stimuli. In the formal hypothesis testing stage, PROCESS Model 83 was applied based on the levels of personalization and privacy concern actually perceived by the participants, in order to test the moderated serial mediation model.
The results show that AI recommendations in social commerce increase consumers’ perceived personalization, strengthen their resonance with the recommended content, and further enhance their happiness, ultimately promoting the formation of consumer loyalty. Resonance and happiness play a significant serial mediating role in the relationship between perceived personalization and loyalty. Additional analyses indicate that privacy concern weakens the positive effect of perceived personalization on resonance and further suppresses the indirect effect of perceived personalization on loyalty through resonance and happiness. These findings suggest that the effect of AI recommendations on consumer loyalty does not simply arise from technical matching efficiency. Rather, it is deeply rooted in consumers’ sense of being understood, psychological fit, and positive subjective experience formed during the interaction process. At the same time, this positive mechanism is significantly constrained by consumers’ perception of privacy risks.
The theoretical contributions of this study can be summarized in four aspects. First, this study extends AI recommendation research in the context of social commerce from a focus on “technical effectiveness” to an explanation centered on “internal psychological mechanisms and relational outcomes.” Second, by incorporating resonance and happiness into an integrated analytical framework, this study clarifies the sequential psychological transmission process through which perceived personalization leads to consumer loyalty, thereby enriching research on the relationships among emotional experience, subjective well-being, and relational behavior in digital consumption contexts. Third, by introducing privacy concern as a key boundary variable, this study further responds to the personalization-privacy paradox and explains why AI recommendations do not consistently enhance consumer loyalty in all situations. Fourth, through a progressive experimental design, this study systematically verifies the core mechanism and boundary condition across different product contexts and variable treatments, thereby enhancing the robustness and causal explanatory power of the findings. Overall, this study provides theoretical foundations and practical implications for social commerce platforms to optimize AI recommendations, balance personalized services with privacy protection, and improve long-term relationship quality with consumers.
Keywords: AI recommendation, social commerce, perceived personalization, resonance, happiness, consumer loyalty, privacy concern.