Abstract

This research examines how an individual’s self-construal and message framing interplay in determining AI recommendation persuasiveness. In study 1, we examine the moderating role of an individual’s chronically accessible self-construal. The results show that interdependent self-construal generated more positive attitudes toward ingroup preference-based recommendation message than individual viewing history-based recommendation message. However, independent self-construal had no significant impact. Study 2 replicates this relationship, but self-construal is temporarily primed. Participants who were primed with interdependent self-construal displayed more positive attitude toward the ingroup preference-based (vs. personal purchase history) recommendations. Participants in the independent self-construal manipulation condition displayed higher preference toward the personal purchase history-based (vs. ingroup preference) recommendations. Our findings provide theoretical implications by extending the previous finding in the context of artificial intelligence recommendation. Also, some managerial implications are offered, especially for marketers and AI technology developers.