Fairness in recommender systems: research landscape and future directions

Abstract

Recommender systems can strongly influence which information we see online, e.g., on social media, and thus impact our beliefs, decisions, and actions. At the same time, these systems can create substantial business value for different stakeholders. Given the growing potential impact of such AI-based systems on individuals, organizations, and society, questions of fairness have gained increased attention in recent years. However, research on fairness in recommender systems is still a developing area. In this survey, we first review the fundamental concepts and notions of fairness that were put forward in the area in the recent past. Afterward, through a review of more than 160 scholarly publications, we present an overview of how research in this field is currently operationalized, e.g., in terms of general research methodology, fairness measures, and algorithmic approaches. Overall, our analysis of recent works points to certain research gaps. In particular, we find that in many research works in computer science, very abstract problem operationalizations are prevalent and questions of the underlying normative claims and what represents a fair recommendation in the context of a given application are often not discussed in depth. These observations call for more interdisciplinary research to address fairness in recommendation in a more comprehensive and impactful manner.

Hybrid Inductive Graph Method for Matrix Completion

“…In the research conducted on recommenders, one area that has recently gained attention is fairness (Giap et al, 2022;Deldjoo et al, 2024). A drawback of recommendation systems is the issue of confirmation bias, which arises when diverse options that might be of interest are blocked, and recommendations are limited mainly to popular items.…”

Consumer-side fairness in recommender systems: a systematic survey of methods and evaluation

“…One such categorization distinguishes between adaptations targeting model input, the model itself and the output. This categorization is well established within the field of recommender system fairness and machine learning as a whole, and the classes are typically named pre-, in-and post-processing, respectively (Caton and Haas 2023;Mehrabi et al 2021;Deldjoo et al 2023). To structure the identified research, we propose a taxonomy based on these classes.…”

“…There has been a recent surge in proposed surveys of fairness in recommender systems. Pitoura et al (2022) surveys both fairness in IR ranking and recommender systems, while Deldjoo et al (2023), Wang et al (2022), Li et al (2023) focus on recommender systems. Pitoura et al (2022) seeks to serve as an overview of fairness in both IR ranking and recommender systems, which makes it the broadest and most high-level survey among the ones considered relevant.…”

“…Because of the broad scope and since it is the oldest survey considered relevant, only two of the studies covered in this survey were covered in Pitoura et al (2022). Deldjoo et al (2023), Wang et al (2022), Li et al (2023) were all first made publicly available within a few months of each other in 2022, and all consider a broad scope comprising all types of fairness in recommender systems. This scope is wider than the one applied in this survey by covering provider-side fairness and group recommendation.…”

Diversity and neocolonialism in Big Data research: Avoiding extractivism while struggling with paternalism

“…This is primarily because economic relevance is often understood as synonymous with smooth market integration based on rapid up-scaling potential (Pfotenhauer et al, 2021). Scaling data sets to promote diversity in automated matching may well be beneficial, such as in creating richer data sets that enable better/fairer performance of collaborative filtering (Deldjoo et al, 2023). However, rapid demands for upscaling are detrimental for a project that seeks to work on par with diverse communities if it is used to replicate the same pilot (ethics, rules, processes, forms of engagement, etc.)…”