AI-driven personalized advertising and consumer purchase decisions: Examining the effects of ad relevance, consumer trust, privacy concerns, and purchase intention

Abstract

AI-driven personalized advertising can simultaneously increase consumer value and intensify concern about how personal data are collected and inferred. This study examines that personalization–privacy tension by testing the effects of perceived personalization, advertising relevance, consumer trust, and privacy concern on purchase intention and purchase decision. A reproducible synthetic dataset representing 320 digital consumers is used for methodological demonstration because no real respondent dataset was supplied. Six constructs were measured with four five-point Likert items. Reliability, Pearson correlation, multiple regression, and interaction-based moderation analysis were applied. Scale reliability ranged from alpha = .882 to .915. Personalization strongly predicted relevance (beta = .601, p < .001). In the purchase-intention model, relevance (beta = .305, p < .001), trust (beta = .294, p < .001), and personalization (beta = .198, p < .001) were positive, while privacy concern was negative (beta = -.124, p = .006). Crucially, the personalization × privacy-concern interaction was negative and significant (beta = -.105, p = .019), indicating that the beneficial effect of personalization weakened as privacy concern increased. Purchase intention strongly predicted purchase decision (beta = .486, p < .001). The findings illustrate the personalization–privacy paradox: AI targeting can improve purchase outcomes, but intrusive or poorly governed data practices can reduce those gains. The simulated nature of the dataset is explicitly acknowledged and real-data replication is required.