The rich get richer and the poor get poorer? The effect of news recommendation algorithms in exacerbating inequalities in news engagement and social capital

Personalized news recommendations shape social media users’ information environment. However, whether news recommendation algorithms asymmetrically influence users’ news engagement remains largely unknown. Drawing on the three-level digital divide framework (access, use, and outcomes), we test a moderated mediation model in which social media usage motivations influence social capital via news engagement, conditional on using algorithmic news. Using two waves of survey data from South Korea ( N = 948), the results show that the indirect effects of motivations for social media use on social capital via news enagement are conditional on the level algorithmic news usage. News algorithms enable information- and socialization-oriented users to increase news engagement and develop social capital but fail to help highly entertainment-focused users increase news engagement, and thus, they do not develop social capital well. We discuss the possibility that news recommendation algorithms lead to a Matthew effect in which the poor become poorer and the rich become richer, exacerbating information inequality.