In recent years, sellers have increasingly relied on artificial intelligence (AI), rather than human employees, to recommend products to consumers. However, if AI-based recommendations elicit negative consumer responses, firms may need to reconsider t...
In recent years, sellers have increasingly relied on artificial intelligence (AI), rather than human employees, to recommend products to consumers. However, if AI-based recommendations elicit negative consumer responses, firms may need to reconsider their AI deployment. The present research proposes that consumers’ purchase intentions in product recommendation contexts vary depending on the type of recommendation agent (human vs. AI), that this effect is mediated by persuasion knowledge, and that the underlying mechanism is moderated according to the brand–consumer relationship type. Across two scenario-based studies, findings demonstrate that AI-based product recommendations activate higher levels of persuasion knowledge than human recommendations, resulting in lower purchase intentions. Moderated mediation analysis reveals that this AI aversion effect emerged when the brand–consumer relationship is communal, whereas no differences in persuasion knowledge activation and purchase intention are observed when the relationship is exchange-oriented. By identifying when and why AI-based recommendations undermine consumer responses, this research offers important theoretical and managerial insights into the effective use of AI in product recommendation settings.