ARTIFICIAL INFLUENCE? A REFLEXIVE THEMATIC ANALYSIS OF AI-DRIVEN PERSONALIZATION AND THE TRUSTWORTHINESS OF ONLINE NEWS

<p dir="ltr">The internet and social media platforms have transformed the landscape of information exchange, granting unprecedented access to diverse news sources and perspectives. To manage this information abundance, news organizations and technology platforms increasingly rely on personalization algorithms to tailor content to individual users (Thurman, 2018). While personalization can enhance content discovery and engagement, persistent concerns regarding their transparency, data use, algorithmic bias, and the potential reinforcement of information silos remain (van Drunen, 2019; McCarthy, 2022; Sehl, 2023), coinciding with rising public skepticism toward the trustworthiness of online news (Fletcher, 2023).</p><p dir="ltr">As AI-driven personalization systems become more prevalent, they may further amplify both these opportunities and vulnerabilities. This study addresses two research questions: (1) How is AI-driven personalization being implemented in the distribution of online news, and in what contexts? and (2) What are the implications of AI-driven personalization for the future of online news? Using a reflexive thematic analysis of 11 peer-reviewed studies published between 2013 and 2023, the research identified cross-cutting patterns related to implementation mechanisms, deployment contexts across platforms and news organizations, audience trust and credibility signals, evolving ethical and governance frameworks, and structural implications for news markets and editorial institutions. Overall, the findings indicate that AI-driven personalization operates less as a discrete feature and more as a pervasive infrastructure that can shape what audiences encounter, how content is presented, and how legitimacy is interpreted within digital news environments. Trust-related responses appear closely associated with transparency, explainability, visible editorial involvement, and opportunities for user agency, while opacity surrounding data practices and ranking logic is linked to heightened uncertainty regarding fairness and credibility.</p><p dir="ltr">Practically, the results suggest that trust-supportive personalization may benefit from clear recommendation disclosures, content provenance indicators, explainable interface designs, and user-adjustable controls, complemented by governance approaches emphasizing shared accountability, interoperability, and proportional oversight. As one of the early focused examinations of AI-driven personalization within journalism, this study provides an interpretive foundation for future empirical research and for ongoing efforts to guide the responsible design and governance of algorithmically mediated news distribution.</p>