I’m presenting some early results of our large-scale dynamic practice mapping of Australian climate change discussions on Facebook later in this next session at the Weizenbaum Conference, but we begin with a paper by Konstantin Lackner, Markus Uhlmann, and Viktoria Horn. Their focus is on news navigation and recommendation: recommendations enable users to navigate information overload, but also create potential monetary gain for content sources.
Recommendations can be problematic because they optimise for retention and attention, and therefore for profit; this is also increasingly done through AI; and the result of such recommendations may be that users no longer encounter the information that may be most important to them.
What might an alternative approach to news navigation look like, then, which maximised for democratic values? Such an approach might be optimised to disclose pluralistic perspectives in news articles, and indeed emphasise argumentative perspectives.
The present project explored this using news articles on electromobility: these articles were labelled for their argumentative perspectives (e.g. environmental or market orientations), and transformed into a conceptual map that connects articles based on shared perspectives. In turn, this conceptual map became the basis for a prototypical system that enabled users to explore news content written from these diverse perspectives, and/or focus on articles presenting highly diverse or highly unified perspectives.
The project then explored user experiences with this system: they experienced stimulation and hedonic qualities, but also found the system unfamiliar and unintuitive; consumption behaviour diverged markedly between users seeking content from only one perspective, and users seeking out a diversity of views.
The project therefore makes the data structure more transparent to its users, and this enables them to engage in autonomous information seeking behaviours; this is different from other models that explicitly identify the supposed political leanings of news sources and provide users with an opportunity to encounter articles from such diverse political perspectives. The present model is more issue-specific in its approach.
