CAIS AG: AI in digital public participation

What role can and should AI play in digital participation processes? These and other questions were discussed by a working group (WG) of DIID researchers and participation practitioners at CAIS in Bochum from February 8 to 10, 2023.

Last year, DIID members Dr. Dennis Frieß and Anke Stoll applied for funding for a working group at the Center for Advanced Internet Studies (CAIS) in Bochum and prevailed in a competitive selection process. From February 8 to 10, the working group, consisting of six DIID AI researchers and five participation practitioners, met at CAIS to exchange ideas on the topic of AI in participation processes.

How and at what point in the participation process can AI-based tools support digital public participation? Can AI potentially make participation more inclusive, clearer and more relevant? What risks are associated with it and is AI at all communicable in the context of democratic policy-making (vis-à-vis politics, but also the addressed public)? What experiences exist in practice with AI applications and how can these be scientifically evaluated? These and other questions were on the agenda.

The working group, which was organized as a workshop, started with a focus group in which the aim was to learn from practice what AI can and should be used for and at what point in the participation process. In addition, various AI interventions were discussed and prioritized. An important finding was that the question of who uses AI should not be neglected. At least three stakeholder groups are relevant here: the users, the implementers (such as facilitators, service providers, etc.) and the commissioners (such as administrations or associations).

Day two was dedicated to science. Here, DIID AI researchers presented three AI-related projects. The first project was the UPEKI project, which aims to use AI to improve the deliberative quality of participation discourses. The so-called KOSMO project, which is primarily concerned with the automated detection of low- and high-quality discussion contributions, focused on AI-supported collective-social moderation of online discourse. Finally, Tobias Escher and Julia Romberger gave insights into their research from the CIMT project – Citizen Involvement in Mobility Transitions. Here, the focus was on the automated evaluation of large amounts of text that arise in participation processes. In addition to the input from the scientific community, the main focus was on feedback and the question of the practical relevance of the various projects.

In three moderated discussion rounds, the aim was to once again systematically identify the needs of AI, anticipate opportunities and risks, and also discuss normative guidelines for the use of AI in participation processes. The findings of the working group are to be systematized again in the coming months and finally published. However, it has already become clear that the dialog between science and practice stimulated by the working group is invaluable, especially for application-oriented research.

The organizers and researchers would like to thank the DIID cooperation partners Liquid Democracy, Polidia and Zebralog as well as ifok (no cooperation partner), who took the time to travel to Bochum. We hope that this exchange will continue. We would like to take this opportunity to thank CAIS for their generous sponsorship and for providing a productive space.