Argument Structures in the Automatic Detection of Intolerance and Extremism

The project "Argument Structures in the Automatic Detection of Intolerance and Extremism" investigates extremism, intolerance and echo chambers on the Internet in contrast with respect and a culture of dialogue. In our era of social media, fake news, shitstorms, increasing fragmentation of society and pandemic isolation, these issues have become more important than ever.

The aim of this project is to recognize trends in digital media and identify prevalent associations, attitudes and opinions. The project will contribute to an understanding of where and how respect is suppressed by ideology. It will focus on the arguments that are the foundation upon which discussion participants construct their opinions about religious matters and other societal issues.

We measure and interpret these factors using content analysis, statistical methods of digital humanities and artificial intelligence. These include text classification, topic modeling, neural networks, distributional semantics, conceptual maps, argument mining and clustering. We will then compare the results of our methods with survey results.

A constructive, open dialogue that is characterized by respect, tolerance and the willingness to engage with one another is the cornerstone of democracy. Ideologies - religious and otherwise - can undermine these values. We plan to make our collected data available to the public in an anonymized form.