“Hey AI, What Should I Eat?” Navigating Skepticism and Trust in AI-Powered Meal Recommendations Through Personalized Persuasion
The Anh Phan, Thi Huong Thanh Nguyen, Cindy Nguyen
Abstract
As AI-powered persuasion agents increasingly enter everyday life, consumer interaction with tools like ChatGPT has extended into domains once dominated by human expertise—including dietary planning. This study examines how personalization, platform credibility, and perceived value influence consumer engagement with AI-generated meal recommendations. Drawing upon the Integrated Fear Acquisition Theory (IFAT), we conceptualize consumer skepticism as a key cognitive manifestation of underlying AI-related fear. Using a lab-based cross-sectional design with direct exposure to ChatGPT, we examined how these three antecedents influence behavioral intention both directly and indirectly through skepticism. The results demonstrate that platform credibility and perceived value significantly enhance user intention and reduce skepticism, while personalization, when operationalized in a minimal form, has no significant impact. Skepticism emerges as a critical mediator shaping consumer acceptance of AI recommendations, particularly in low-risk, health-oriented contexts among young adults.