As recommender systems take the editor’s seat, many ask: Why do we still have editorially managed public service media? From the recent discussions of fake news and societal cohesion, it seems however less straight forward to replace editors with algorithms. Algorithms may be easy to trick and bias, and in most cases the inner mechanics of their operations are not revealed to the public. However, in order to remain competitive and attractive also public service media organisations implement recommender systems. The job is not easy as special requirements apply to public service media: They can not all allowed accusations of being biased in recommendations. Furthermore, the need to live up to a number of obligations regarding publishing output - typically defined by media politicians.
In spring 2019, the CMI researcher Jannick Kirk Sørensen met in Hamburg at Hans-Bredow-Institute for Media Research (Hamburg University) to discuss with international colleagues how public service media can react to and navigate in the world of algorithmic recommendations. The question has for the last couple of years been in focus at CMI as part of the ‘Digital Media’ area. We thus continue to examine the question: How does a recommender system for a public service media look like? How to ensure unbiased diversity in its recommendations? What is the role of publicly funded media in the age of algorithms? How big is the user’s say in relation to algorithms that governs which news and media content they see? What about privacy and the trustworthiness of public service media? These questions calls for more research.