Author Name: Ashish K Date: 28-03-2026
Algorithmic bias in news recommendation systems — systematic patterns of error or distortion in how recommendation algorithms treat different user groups, content types, and journalistic topics — produces psychological consequences that fall disproportionately on already marginalized audiences. This paper provides a comprehensive analysis of the sources, manifestations, and psychological impacts of algorithmic bias in news recommendation, integrating algorithmic fairness research, media psychology, and critical race and gender studies to develop the first comprehensive theoretical framework for equitable news algorithm design. The paper identifies four sources of algorithmic bias: historical data bias (training data reflecting existing inequalities in news production and consumption), representation bias (underrepresentation of marginalized communities in model development), measurement bias (performance metrics that optimize for majority user behavior), and feedback loop amplification (algorithmic reinforcement of biased patterns through engagement-based learning). The psychological consequences for marginalized audiences are analyzed at three levels: identity-level (reduced psychological representation and recognition); epistemic-level (systematic underexposure to news relevant to marginalized communities and perspectives); and wellbeing-level (hostile media perceptions, alienation from journalism institutions, and increased news avoidance). The paper proposes an Equitable News Recommendation Framework (ENRF) that specifies minimum fairness standards for coverage equity, source diversity, and content representation, and proposes algorithmic audit protocols for ongoing bias monitoring. The consequences of uncorrected algorithmic bias for democratic representation and media trust among underserved communities are analyzed.
Keywords: algorithmic bias; news recommendation; marginalized audiences; representation bias; media equity; algorithmic fairness; news avoidance; democratic communication.
