“I must have clicked on something”–Users´ Experiences and Evaluations of News Recommender Systems

This article takes a user perspective on AI-driven news recommender systems (NRSs). We argue that understanding users’ experiences with recommender systems is crucial to assess the implications of AI and automation for the perceived legitimacy of news media in democratic societies. Much of the literature on users and NRSs have either focused on measuring user behavior in controlled set-ups or on the user’s imaginaries of personalized recommendations as a phenomenon. In this article we employ a novel methodology for making NRSs tangible by exposing 24 participants to personalized recommendations based on their news use, interviewing them first individually about their experiences and evaluations, and then collectively about the implications of personalized news recommendations. The analysis shows that users make sense of personalized recommendations in relation to personal relevance, the news organization and its business model, and the broader media ecology permeated by algorithmic curation and filtering. Importantly, these findings suggest that personalized news brings about a shift in users’ perception of their role on a news site; they are not only participants who can shape their own news experiences, but also actors who are co-responsible for the content they are exposed to.