Abstract. Spread of hatred online is increasing and can be related to serious psychological and criminal outcomes. The goal of this project is to identify patterns of interaction that lead to a decrease of cyberhate. We call this kind of interactions counterspeech. To quantify the effect of counterspeech on cyberhate, we have collected a larger number of tweets and their replies from German Twitter accounts that are often targeted by hate accounts. The unique situation in Germany is that there are at least two opposing groups active on social media. One is spreading cyberbate (Reconquista Germanica) while the other was formed as a counter movement (Reconquista Internet). This gives us the unique opportunity to study the effects of organised counterspeech on organised cyberhate. The first step in this project is to build a system that is able to detect cyberhate and counterspeech and distinguish them from other tweets.
In this talk, we will present the background that lead to this project, the data that we have collected, the scientific questions that we are asking, and finally, our first steps to automatically identifying cyberhate and counterspeech.