Mediacenters of the Future

In the age of digital media available on the internet users have access to abundant content and need to deal with a vast amount of information. Broadcasters and other media companies are searching for solutions to provide users with relevant content. Primarily technology-driven commercial media providers such as YouTube, Netflix or Amazon use personalization and user profiles for context and usage specific recommendations.

Balancing diversity and personalization in the presented content of the video platform becomes a central issue when it comes to public broadcasters. These institutions face the challenge that they cannot pursue strategies that only maximize the economic success of their digital content. In Germany they should rather take into account “the principles of objectivity and impartiality of reporting, the plurality of opinions as well as the balance of their offerings” according to the German Interstate Broadcasting Agreement. Trying to accomplish these competing goals – serving users’ individual interests to create user loyalty and identification with the service on the one hand and fulfilling the duties from the Interstate Broadcasting Agreement on the other hand – has consequences both for the use and perception of the content by recipients and the work in editorial and development departments.

We examine how this socio-technical design task can be solved in the context of the prototype development for a specific platform for on-demand video and audio content. The study focuses on the interfaces between the technical platform, editors and users and explores in an interdisciplinary team, including engineers and editors, ways to shape a data driven and algorithmic system of content mediation that meets the democratic and social obligations of public broadcasters.

Using collaborative ethnography of the software development process and the involved information infrastructures as well as a qualitative performance analysis of different recommendation algorithms we intend to show how different platform strategies influence production and diversity of content.

For media and internet politics a central issue is what effects these strategies have on users’ perception of content and the co-production of media publics. Seemingly neutral ranking and recommendation algorithms comprise built-in preferences whose mechanics users rarely understand and which users can hardly influence. Yet, personalization and prediction algorithms already have a central role as mediators of media content and this kind of software will become even more important in the near future. In exploring ways to shape these algorithms we also explore and negotiate newly emerging media ecologies and publics.