In June 2022, an unusual democracy experiment was running: randomly selected citizens debated the consequences of the use of artificial intelligence in a virtual citizens' assembly. The content was about the use of self-learning machines in nursing and care. Whether and in what way algorithms should assist human resources departments in their hiring decisions was also a topic.
Possible questions were: Will artificial intelligence make personnel managers superfluous in the search for new skilled workers? Or can it be an advantage to achieve better results in the recruitment process through additional data processing? Do we really want to discuss personal and medical problems at the end of life - instead of with a human caregiver - with a robot that, for example, has the appearance of a plush seal that you hold in your arms?
200 people took part in the Citizens' Assembly in a virtual discussion group and 100 more in a control group. The Citizens' Assembly members came from all over Germany. In order to capture the effect of the exchange in the Citizens' Assembly, the control group only received the arguments on the assembly topics, while the discussion group could also discuss the arguments. The assembly took place "asynchronously" on ten days. The participants were therefore not online at the same time. They had ten time slots to log in to the participation portal "Demokratiefabrik".
According to its own statement, the participation portal guarantees that a maximum diversity of arguments and positions is included in the discussion. Assembly members can write their own contributions as well as evaluate and comment on randomly selected contributions from other participants. In addition, four experts help the participants in finding their opinions.
Students from the University of Stuttgart and the Stuttgart Media University facilitated the Citizens' Assembly. Participants were randomly assigned pseudonyms. In this way, the arguments put forward should carry more weight than the social background. In addition, the exchange of arguments and the interaction with experts could be observed and analysed.
"Artificial intelligence is a central topic of the future"
Project leader Prof. Dr. André Bächtiger, Executive Director of the Institute for Social Sciences at the University of Stuttgart, hoped to gain new insights into the democratic process of debating in this way. "Artificial intelligence is a central topic for the future. In our project, we are investigating how it can be possible to systematically involve citizens in challenging normative questions," said Professor Bächtiger. Prof. Dr. Alexander Mäder, who was supervising the project at the HdM, emphasised: "Today's students are tomorrow's science communicators. Participatory processes are becoming more and more important here."
The aim of the experiment was to be able to pass on recommendations or at least problems to politicians and society that are important for the future regulation of artificial intelligence through questions posed by the participants, discussion with experts and moderation by the students.
Citizens' Assembly part of the "Questions to Colleague AI" project
The democracy experiment was part of the project "Questions for the colleague KI", a project of the University of Stuttgart (Interchange Forum for Reflecting on Intelligent Systems, IRIS, International Centre for Cultural and Technological Research, IZKT) and the Stuttgart Media University in cooperation with the Stuttgart Public Library. Among other things, the two universities are organising a teaching project on participatory science communication and a series of public debates. The project was funded by the Federal Ministry of Education and Research (BMBF) as part of the "Year of Science 2022 - In Demand!".
The results of the Citizens' Assembly on Artificial Intelligence will be published in autumn 2022.
Read more: Demokratiefabrik