Is Algorithm the Culprit of Information Cocoons: A Systematic Literature Review

This study systematically reviews the relationship between algorithmic recommendation and information cocoons. Using Biggs’ 3P model, it classifies key elements into user characteristics, technological environments, user–algorithm interaction processes, and outcomes, and proposes four research questions. Following PRISMA 2020 guidelines, English empirical studies published between 2010 and 2026 were searched in four databases. After screening, 15 studies were included. Results show that information cocoons are shaped by users’ political leaning, preferences, usage habits, and interaction patterns, reinforced by feedback loops, information homogeneity, and immersive interfaces, and partially mitigated by diversification-oriented mechanisms. User–algorithm interactions influence this process through direct reinforcement and indirect mechanisms such as filtering and immersion, while information cocoons produce cognitive, behavioral, and social consequences. The study highlights limitations and suggests directions for future research.