Logo of the DeLab project. Since 2021, I’m coordinating the Deliberation Laboratory (DeLab) project. The project is funded by the Volkswagen Foundation within their funding line Artificial Intelligence and the Society of the Future. For more detailed information about the project please visit the project website: – here I will only provide a brief sketch of the project:

It is well known that unmoderated online discussions on social media rarely stay civil. On the contrary, we see phenomena like hate speech, reasoning fallacies, echo chambers and fake news. It is also well known that active moderation of social media discussions can improve the quality of such discussions, bringing back (some sort of) civil behavior. However, active moderation is rather a hard and expensive task: For instance, with about 500 million tweets per day, moderation means “deletion” and nothing more complex than that. What if, instead of human moderators, a virtual moderator enters the scene in real time? Imagine a virtual moderator that can detect when things go wrong in social media discussions, and intervenes on its own to steer discussions towards a more respectful debate.

This is exactly what ultimately aim for in the DeLab project: Building an artificial intelligence moderator intervening on its own to improve the quality of social media discussions. There are, of course, lots of questions involved and we strive for answering at least some of them. That is the “lab”-part of the project: Before introducing the moderator in the wild, we extensively test the moderator to rule out unintended consequences. AI might help in some instances, but fail to do so in others.

Besides the project webpage, you’re also invited to read the official press releases by the Georg-August-Universität Göttingen (German, English). DeLab has also been featured at NDR (Northern German Broadcasting) in the news show Hallo Niedersachsen.