To what extent can algorithmic recommendation systems be part of the own journalistic activity of public service media providers and take the side of (or take the place of) journalistic selection and compilation of information? In a White Paper for the MDR (Central German Broadcasting), the main characteristics and principles of algorithmic recommendation systems are summarised and their consequences for public service offers are discussed.
The increasing significance of digitally networked media has fundamentally changed the structures of the social public sphere. An essential part of this change is that the selection, processing and presentation of information is no longer carried out by the editorial offices of journalists and publicists alone.
Especially with search engines like Google or social media intermediaries like Facebook, YouTube or Twitter, algorithmic selection and recommendation systems perform the indispensable task of selecting from the wealth of information and content available to users. An essential performance promise of algorithmic systems in this context is the “personalisation” of information offers, so the most individual compilation of relevant, interesting and otherwise suitable recommendations. For some years now, however, it has also been critically discussed whether algorithmic personalisation may contribute to social fragmentation and polarization, using keywords such as "filter bubble" or “echo chamber”.
This development is forcing media organisations to consider strategically to what extent algorithmic recommendation systems should be part of their own journalistic activity, meaning to what extent they should take the side of (or take the place of) journalistic selection and compilation of information. For public service broadcasters, their constitutional mandate raises special questions, such as the balance between personalisation on the one hand and basic provision on socially relevant issues on the other hand.
Against this backdrop, the Hans-Bredow-Institut has produced a White Paper for the MDR summarises the main characteristics and principles of algorithmic recommendation systems and discusses their consequences for public service offerings.
The White Paper has been published as 'Working Paper of the Hans-Bredow-Institut, No. 45' (in German):
Jan-Hinrik Schmidt / Jannick Sørensen / Stephan Dreyer / Uwe Hasebrink (2018): Algorithmische Empfehlungen. Funktionsweise, Bedeutung und Besonderheiten für öffentlich-rechtliche Rundfunkanstalten [Algorithmic Recommendations - Functionality, Meaning and Peculiarities for Public Service Broadcasters]. Hamburg: Verlag Hans-Bredow-Institut, September 2018 (Working Papers of the Hans-Bredow-Instituts, No. 45) (pdf)
A short version of the white paper has been published as article in the journal MediaPerspektiven (in German):
Jan-Hinrik Schmidt / Jannick Sørensen / Stephan Dreyer / Uwe Hasebrink (2018): Wie können Empfehlungssysteme zur Vielfalt von Medieninhalten beitragen? Perspektiven für öffentlich-rechtliche Rundfunkanstalten. In: Media Perspektiven, 11/2018, S. 522-531. (pdf)
