iconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsiconsicons
Members of the Institute Represent German Communication Science at the ICA 2018

Members of the Institute Represent German Communication Science at the ICA 2018

At the 68th Annual Conference of the International Communication Association (ICA) in Prague, the panel of the German Society for Journalism and Communication Studies (DGPuK) (LINK) took place on the morning of 28 May 2018, in which German communication science presents itself annually with a main topic to the international specialist community. This year, the proposal of Julius Reimer from the Hans-Bredow-Institut was selected in a peer review process. The proposal “Opening the Black Box: Investigating the Algorithmization of Journalism” included four lectures dealing with the increasing influence of algorithms in journalism from different perspectives.

In the first lecture of the panel, Wiebke Loosen, Stephan Dreyer and Amélie Heldt first classified the field and gave an overview of the four different forms of use of algorithms in journalism:
  1. Algorithmic journalism, in which journalistic content is distributed by personalisation and search algorithms selected via search engines, social networks and other online platforms;
  2. Metric-driven journalism, which uses algorithmically produced click numbers and other measures values to observe the reactions of the audience to contributions and in which, in extreme cases, topics are selected and designed in such a way that personalisation and search algorithms give preference to them;
  3. Data journalism, in which algorithms are also used to evaluate algorithmically generated data and present it in journalistic articles; and
  4. Automated journalism, in which algorithms independently produce journalistic articles based on structured data sets. (Systematisation was developed by Wiebke Loosen, also available in this working paper)
In the subsequent analysis of possible legal consequences of these developments, the three researchers pointed out, among other things, that media organisations in Europe must ensure that they continue to meet the requirements necessary to retain the legal privileges of a journalistic medium when using algorithms.
 
In the second lecture, Julius Reimer examined together with Folker Hanusch, professor for journalism at the University of Vienna, and Edson Tandoc, assistant professor at Nanyang Technological University in Singapur, how the metrics mentioned above affect the ideas that journalists have of their audiences as well as their interests and attitudes. Using empirical data from the project of the Hans-Bredow-Institut on the "(Re-)Discovering the Audience“, they showed that metrics seem to actually bring about changes. At the same time, however, the non-algorithms-based user comments on websites and social media profiles as well as more conventional information sources such as reader’s letters to the editor, sales figures, audience ratings and personal encounters also have an impact – and this in different ways. The researchers concluded that the focus only on metrics and their influence is not sufficient. Instead, studies are necessary that – similar to the research on media use at the HBI – focus on the entire repertoire of information sources, so that journalists get an idea of their audience.
 
The lecture of Sascha Hölig and Lisa Merten shifted the attention from journalists entirely to the audience and its view of the algorithmic journalism mentioned above. Based on representative survey data from the Reuters Digital News Survey and qualitative interviews with media users from another project of the Hans-Bredow-Institut they showed that the vast majority of people do not have an exclusive preference for message selection only by algorithms or journalists but prefer a combination of both options. Both forms of selection would be attributed different advantages and disadvantages, and thus combined users would also be consciously offered media that stand for one or the other type of selection in order to meet their different needs and to be fully and diversely informed.
 
In the fourth lecture, Carl-Gustav Lindén and Hannu Toivonen of the University of Helsinki focused on the fourth and probably most pronounced form of algorithmisation in journalism: automated text production. They gave an insight into the inner life of Bot ‚Valtteri‘ developed by their team, which automatically produces articles on election results for any constituencies and candidates in three languages “at the push of a button”. The system is currently being modified for automated crime reporting. Communication and computer scientists paid a particular attention to the limitations of automated text production. Thus, the underlying principle has not developed decisively for years. The much cited image of the "robot journalist", who will replace human reporters in the near future, is therefore also wrong and hopes and fears currently expressed by journalists regarding the spread of automated journalism are exaggerated. Instead of using a robot for comparison, automated text production is more comparable to a washing machine. Above all, it does tiresome tasks that used to be done by hand, increases productivity, and nobody would seriously wish for a return to the time before its invention.
 
In a response to the four lectures, Natali Helberger, professor for Information Law at the University of Amsterdam and a proven expert on the algorithmisation of media and communication, reunited the different perspectives. For research as well as for the social debate on the increasing datafication and automatisation of journalism and public communication as a whole, it is crucial that the perspectives gathered in the panel are considered more closely together. Different questions arose, for example whether journalists are better than algorithms in any case, since people also tend to make unilateral decisions. Nor is it clear how to empirically measure and normatively evaluate the diversity of opinions and information in order to direct algorithms towards their maximization, nor is a maximum of diversity of opinion and information actually desirable in view of the social context. For the future, however, she is optimistic that algorithms and people will increasingly complement each other – for the individual benefit of journalists and users as well as for the benefit of society.

The event report was written by Julius Reimer.

Newsletter

Subscribe to our newsletter and receive the Institute's latest news via email.

SUBSCRIBE!