AI-Driven Recommendation Systems and Human Cognition: A Systematic Review of Algorithmic Personalization in Digital Information Ecosystems

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Recommendation systems now sit at the centre of most digital platforms, quietly shaping what users see and click, and potentially how they form beliefs over time.As these systems have grown more sophisticated, driven by advances in machine learning and behavioural data collection, a critical question has gained urgency: how does sustained exposure to algorithmically curated content shape human cognition itself?Research on recommendation algorithms and research on cognitive processing have largely developed along separate tracks, with few attempts to bring the two together into a coherent framework.This paper addresses that gap through a systematic review of literature spanning Artificial Intelligence, Human-Computer Interaction, Cognitive Psychology, Behavioural Science, and Information Science.Rather than starting from the assumption that personalization is either a benefit or a harm, the review takes a neutral, evidence-first stance, tracing how recommendation systems influence attention, information exposure, and decision-making, and how, in turn, user behaviour reshapes the very algorithms that serve them.This bidirectional relationship is developed into a conceptual framework describing the feedback loop between adaptive systems and human users.The review also maps out where current research falls short, and considers what these gaps mean for responsible AI design, media and algorithmic literacy, and the broader goal of building recommendation technologies that work with human cognition rather than around it.