• Funding : Université de Fribourg
  • Start year :
  • 2025

Joint work with University of Fribourg

In recent years, technological progress has exposed our democracies to complex challenges, for instance by facilitating misinformation and manipulation. But research around e-democracy has shown that technology can also provide tools to deal with these challenges. One such tool that enjoys great popularity in Switzerland is Smartvote, which provides users with voting advice in elections. So-called voting-advice-applications enjoy great popularity throughout Europe. However, in the Swiss political context, elections are only one important way for the public to make political decisions. Much more frequent and at least as important are popular votes on initiatives and referendums. And online in general, voting on issues is become increasingly popular. Despite that, there exists no comparable advice application for popular votes. This project aims to fill this gap by working towards voting-advice-applications for popular votes.

Thus, the main question this project aims to answer is: How can we generate voting advice for voters in referendums or initiatives?

Current voting-advice-applications take the views of a voter and compares them to those of the candidates in order to find the best match. Since popular votes do not involve candidates, this methodology cannot be straightforwardly adapted. Instead, this project takes the following approach: A voting decision should be based on a comparison of the arguments that speak in favour of the proposed state action (pro-arguments) and those that speak against it (con-arguments) and thus a voting advice mechanism should evaluate and weigh these arguments from the perspective of a voter in order to decide whether voting “yes” or “no” would be best for them. This approach situates the project mainly in argumentation theory and decision theory, taking inputs from both philosophy and computer science. The aim is to work towards a computational framework that is philosophically well-founded.