Like virtually all social domains, political communication is profoundly impacted by the deployment of big data. In this paper we focus on one case of this big data revolution: the rise of algorithmic news recommenders (ANR) in a changing media ecosystem. More specifically, we discuss the impact of personalized news media on the formation of a shared agenda of the public discourse. Many studies point to the growing role of ANR as means to specifically tailor news items to individual users. This allows media to adapt their news provision to specific user preferences, and empowers news readers as they can bring up issues beyond dominant news stories. News personalization could, therefore, make citizens more engaged in public debates. At the same time, the rise of ANR can have disruptive influences on the public sphere by leading to selective exposure to news which diminishes the possibilities of shared public debates. Ideological polarization together with so-called filter bubbles and echo chambers are often assumed to be the result of ANR.
We contribute to the ongoing debate about the impact of ANR on the public sphere by focusing on the personalization practices used by digital outlets of traditional news media such as newspapers and TV-channels. The majority of existing studies on agenda-setting and news personalization focus on social media such as Twitter and Facebook, whereas the implications of algorithmic targeting and profiling by mainstream news media remain under-investigated. To address this subject, we use a conceptual approach to operationalize the interactions between recommender algorithms, editors, and users in the context of agenda-setting. Among the questions that we address are the following: Who are the new gatekeepers in the algorithmically infused media landscape? What is the role of human agency in the data-driven process of news personalization? And ultimately, how are power configurations in the domain of news agenda-setting changing due to the rise of ANR and how it affects free and non-discriminatory access to information?
