Digital Deliberative Democracy

Democracy is seemingly in turmoil: the legitimacy of won elections is doubted due to interference via social media; public ballots and referendums are going differently than expected; and the deliberative culture within countries is seemingly disrupted by the increasing polarization supposedly fostered by the rise of social media and their lacking robustness against manipulations. The rise of novel electronic media seems to destabilize our democracies. Does that mean that the wisdom of crowds that democracy taps via its participatory mechanisms gets drowned in the cacophony of off-topic or biased posts, lost in the echo chambers, and manipulated (or ‘nudged’) beyond usefulness?

There are reasons to be optimistic: novel electronic coordination approaches, via social media, email, instant messaging, or discussion platforms, have also been shown to raise civic participation. A plethora of civic tech platforms try to harness the power of online participation. What these tools have in common, though, is that they leverage electronic communication to extend participation without understanding the possible pitfalls mentioned above. Even more so, they often do not take into account insights from research on online collaborative work or collective intelligence on how to build systems that systematically harness the capabilities and cognitive variability of large-scale crowds. Finally, these tools are often not embedded in the political process and provide few practical insights.

Pipeline for Online-Participation

To that end, this project will investigate how technology-supported, large-scale crowd computing approaches can be used to strengthen the functions of the Swiss consultation and popular initiative procedures. These two procedures are arguably the most impactful participatory mechanisms of direct democracy, as they allow the population to not only strongly influence the outcome of the political process but also to set the agenda (in popular initiatives). Whilst they look different from the outside, they share significant structural similarities: Both are complex mechanisms, as they require

1. the identification of the controversial issues;
2. a deliberation to find a suitable solution as well as to assess its material accuracy, feasibility of implementation, and public acceptance;
3. the refinement of the deliberation’s result into a submission document.

Hence, this project will build novel approaches – essentially hybrid human-machine systems that cultivate, coordinate, and support participants using coordination technology and artificial intelligence (AI) – to support these phases and test them in four real-world settings: the classroom, the Swiss and German democratic setting, and the Wikimedia community. These four settings allow us to hypothesize, test, and refine political science theories, crowd computing approaches, and legal frameworks in increasingly complex and uncontrolled settings, which require subtle and careful consideration of the possible negative effects.

Indeed, given that the project proposes to test the approaches by preparing and submitting response-documents to at least five real-world consultations, it will need to carefully consider its possible interaction with the political realities and its impact. Therefore, it will continuously evaluate all the output in terms of

1. procedural quality (how well did the process run),
2. legislative quality (what is the quality of the generated result),
3. participant satisfaction (are the participants happy with the process & results; will they likely participate again)
4. the democratic legitimacy/acceptance (how likely will the resulting output be acceptable by the legislature and the population at large).

Perfect Playground

Note that Switzerland, due to its long tradition of both federal and cantonal participation, is uniquely suitable to test such tools under meaningful, real-world circumstances. Given that these goals need to combine and advance our understanding regarding the political process, the legal framework, the legislative process, as well as modern crowd-computing and AI techniques, this project will have to combine insights and methods from political science, jurisprudence, AI, and computer supported cooperative work. Through this unique and inter-disciplinary combination our project is expected to

1. uncover what mechanisms help large-scale issue identification, deliberation, and solution design within democratic processes whilst mitigating possibly negative effect of its online nature,
2. help us understand how our current democratic processes could be reformed to increase the quality of their decision making and their robustness against manipulation,
3. provide a novel platform to support democratic decision making,
4. advance science in all participating disciplines.

The insights of this project may prove to be crucial for a direct democracy like Switzerland (and online communities that have adopted similar mechanisms) at this point, to foster democratic innovation and turn its citizens to even more active, AI-empowered participants in the democratic process-a goal which ultimately ensures democratic stability and our welfare.