Purpose Businesses are increasingly using artificial intelligence (AI) to provide recommendations to consumers. However, in many domains, consumers prefer human customer service and demonstrate aversion to AI-driven interactions. This study aims to examine consumers' willingness to accept AI vs human recommendations based on the sensitivity of the personal information needed to access services. Design/methodology/approach We simulated scenarios that consumers frequently encounter – namely, movie ticket purchasing, financial services, online shopping and fitness and/or weight management. Study 1 primarily examined the main effect (i.e. how contextual sensitivity influences users’ willingness to accept different types of recommendations) and verified the mediating role of self-awareness. Study 2 investigated how privacy concern and need for uniqueness moderate the relationship between contextual sensitivity and recommender type when shaping users’ acceptance intentions. To enhance the robustness of the findings, Study 2 also retested the main and mediating effects of the model. Findings In high-sensitivity contexts, consumers tend to favor AI agents over human ones, but the opposite is true for low-sensitivity scenarios. Furthermore, this study investigates the moderating effects of consumer characteristics (i.e. the need for uniqueness and privacy concerns) and the mediating role of (public and private) self-awareness in shaping these preferences. Originality/value The findings offer valuable insights and carry significant implications for corporate decision-making regarding the deployment of AI technologies in consumer-facing services.
Wang et al. (2026) studied this question.