Balancing Personalization and Public Values: Legitimacy and Design of Algorithmic News Recommendations in Public Service Media

This paper addresses the normative challenges of algorithmic recommendation systems in public service media (PSM), proposing a framework that aligns democratic mandates with digital engagement via user needs categorization. Institutions like Radio France and the BBC use algorithms to counter news avoidance and filter bubbles through civically weighted content and discoverability features. Yet, their non-commercial missions create tensions between user autonomy, transparency, and societal value. The study introduces a paradigm shift: modeling user behavior through why-oriented needs (e.g., civic awareness, informational gaps) rather than how-focused engagement metrics. A 2023 case study with Switzerland’s RTS develops three nudging scenarios using community detection and needs-based clustering. These inform a tripartite legitimacy framework: 1) alignment with public service duties, 2) ethical user guidance, and 3) systemic risk mitigation via participatory design. Findings show that needs-aware systems require multidimensional profiling balancing explorability and explainability, distinct from commercial logic. Bridging data science and nudging ethics, this work advances interdisciplinary strategies for operationalizing PSM’s dual mandate: respecting individual agency while fostering democratic citizenship.