Re-centering the user in recommender system research: Preface to the special issue

Recommender systems increasingly shape what people watch, read, buy, and learn, yet their success has too often been assessed through system-centric objectives and proxy metrics. In this editorial introduction to the Special Issue on “Re-centering the User in Recommender System Research,” we argue for a sustained shift toward user-centered, value-sensitive recommender research that foregrounds user agency, transparency, fairness, privacy, and long-term welfare, while acknowledging the multi-stakeholder nature of many recommendation ecosystems. Building on perspectives from recommender systems, human–computer interaction, behavioral science, and responsible AI, we first synthesize key conceptual and methodological developments that motivate placing users at the core of design and evaluation. We then introduce the papers collected in this issue, which collectively examine the benefits and risks of persuasive explanations, frameworks for designing effective explanatory interfaces, user-calibrated approaches to mitigating popularity bias, preference conceptualization and modeling in specific domains, user experience in conversational recommenders powered by large language models, and methodological lessons for evaluating group recommenders. We conclude by outlining open challenges–including measuring long-term outcomes, balancing personalization with risks of manipulation, and developing ecologically valid, stakeholder-aware evaluation protocols–and by pointing to promising directions for future research.