Human or AI? The role of social approval source and perceived privacy protection in augmented reality shopping

Purpose Emerging technologies like augmented reality (AR) and artificial intelligence (AI) are reshaping fashion retail. This study investigates how the source of social approval (human vs. AI) influences purchase intention in AR shopping, focusing on the mediating role of decision confidence and the moderating role of perceived privacy protection. Design/methodology/approach A series of four scenario-based experiments was conducted to test the proposed model. Studies 1 and 2 established the baseline effects and mediation. Studies 3 and 4 employed factorial designs manipulating approval source and privacy protection levels to examine boundary conditions, employing rigorous controls for algorithm aversion and individual traits. Data were analyzed using ANCOVA and moderated mediation analysis. Findings Results indicate that human-generated social approval significantly increases consumers’ purchase intention compared to AI-generated approval. Decision confidence mediates this relationship, demonstrating that human-generated approval enhances consumers’ certainty in their purchase decisions. Additionally, perceived privacy protection moderates this mediation; human-based approval has a stronger positive effect on purchase intention when consumers perceive privacy protection as effective. Originality/value Unlike the Computers As Social Actors (CASA) paradigm or source credibility frameworks, which imply that sufficient social cues or expertise can equalize influence, this study identifies a specific social validation deficit for AI in identity-relevant contexts. It establishes perceived privacy protection not merely as a safeguard but as a critical compensatory mechanism that substitutes for interpersonal trust, thereby extending social influence theory in the AI-AR domain.