AI PERSONALIZATION AND PURCHASE INTENTION: THE MEDIATING ROLE OF CONSUMER TRUST AND THE MODERATING ROLE OF PRIVACY CONCERN

Abstract

Within social media environments that use artificial intelligence (AI) to personalize advertisements, recommendations, and content, the influence of personalization on purchase decisions is uneven, shaped by consumers’ trust in the platform and their concerns about privacy. This study examined the effect of AI personalization on purchase intention, the mediating role of consumer trust, and the moderating role of privacy concern on the trust–purchase intention relationship. Data were collected from 112 social media consumers in Punjab, Pakistan, and analyzed using partial least squares structural equation modelling (PLS-SEM) in SmartPLS 4 with bootstrapping (5,000 subsamples). AI personalization significantly increased consumer trust (β = 0.617, p < .001) and had a smaller but significant direct effect on purchase intention (β = 0.264, p = .008). Consumer trust significantly increased purchase intention (β = 0.552, p < .001), and the indirect effect of AI personalization on purchase intention through consumer trust was significant (β = 0.340, p < .001), indicating partial mediation. The interaction between privacy concern and consumer trust was not significant (β = −0.014, p = .827), and the conditional indirect effect did not differ meaningfully across levels of privacy concern; the moderation and moderated-mediation hypotheses were therefore not supported. The model explained 38.0% of the variance in consumer trust and 57.0% of the variance in purchase intention. PLSpredict indicated established predictive relevance for both endogenous constructs, with PLS-SEM producing lower prediction error than a linear-model benchmark for eight of nine indicators, and the Cross-Validated Predictive Ability Test (CVPAT) confirmed that the model significantly outperformed the indicator-average benchmark. Together, these results suggest that AI personalization is more strongly and reliably linked to purchase intention through the trust it builds than through its direct effect alone, and that digital platforms and advertisers benefit more from trustworthy, transparent personalization than from targeting accuracy alone.