Artificial intelligence recommendations: evidence, issues, and policy

Abstract

Recommender systems (RS) enhance user experiences by providing personalized content and are widely used by popular services like Apple Music, Spotify, Netflix, and YouTube to increase user engagement. However, these systems can also have significant economic implications, including exacerbating market concentration and reducing content diversity. This paper reviews recent economic literature on RS, emphasizing their dual role as both beneficial tools and potential sources of market distortion. The paper underscores the necessity for policies informed by economic research to balance the benefits of RS against their associated risks.