This article theorizes how intelligent recommender systems in digital libraries reshape scholarly reading, information trajectories and the spatial–linguistic configuration of scholarly communication. The aim is to conceptualize algorithmic mediation as a multi-level process linking individual reading practices with global hierarchies of visibility.
The study employs an integrative literature review and critical analysis of international and national research (primarily 2019–2025) on digital libraries, recommender systems, information behavior, academic social media (ASM) and global scientific inequalities, synthesizing them into a multi-level conceptual framework.
The review indicates that recommender modules have an ambivalent impact: at the micro level they shape early-career researchers' identities and cognitive “fields of view”, while at the macro level they project language, regional and popularity biases into visibility asymmetries. The study synthesizes these findings into a three-level model (micro–meso–macro) that explicates algorithmic feedback loops linking user behavior, platform architectures and evaluation regimes governing the global distribution of scholarly attention.
The article contributes a theoretically grounded account of algorithmic mediation in scholarly reading that bridges literature on digital libraries, recommender systems, information behavior and global scholarly inequalities. The proposed model offers a transferable analytical lens for future empirical investigations of recommender systems in academic contexts.
