Trusting the Machine: Consumer Perceptions of AI Versus Human Recommendations

Abstract

This study investigates how consumers perceive and trust artificial intelligence based recommendation systems compared with human recommendations across decision contexts. Using a systematic literature review method, the article synthesises empirical evidence on perceived competence, impartiality, empathy, transparency, learning capability, and privacy concerns linked to artificial intelligence driven advice. The review identifies trust as a multidimensional judgement that combines beliefs about technical performance with inferences about benevolence and integrity. The findings show that consumers may value artificial intelligence recommenders for their efficiency and perceived objectivity, yet often experience them as opaque and threatening to personal control and privacy. Algorithm aversion emerges when visible errors lead consumers to penalise artificial intelligence more harshly than human advisors, although demonstrations of learning and clear explanations can partially restore trust. Overall, the review concludes that willingness to follow artificial intelligence recommendations depends on how systems are governed and communicated, and on whether human warmth and fairness are seen as necessary in the specific decision domain.