Identification, detection, and mitigation of covert propaganda attacks via online media/PropStop




Social and political developments are discussed and commented on worldwide on social media. In addition to the desired public, anonymous and free expression of opinion, however, these services also offer scope for controlled manipulation. Semi-automated to fully automated systems known as propaganda bots or social bots use the easy access to the technical infrastructure of social media to covertly disseminate certain opinions and boost public perception in a one-sided manner. To this end, sophisticated artificial profiles are created, fake discussions are generated between users, real opinions are undermined, and unwanted criticism is suppressed. The distorted image of the online public in turn influences social debates and can also influence media coverage, resulting in significant social damage.


Throughout the project an interdisciplinary team of statisticians, communication scientists, IT-security researchers, journalists and IT-security companies works together. Besides a large-scale examination of propaganda characteristics online, huge quantities of public opinion statements in different areas of the digital public sphere will be analyzed for repeated semantic and technical patterns. The insights gained throughout the project will be used to enhance our detection of massive, hidden propaganda attacks, to develop technologies to identify these attacks, and to improve our abilities to verify propaganda attacks. Real-time simulations of massive hidden propaganda attacks will provide meaningful insights into the transdisciplinary applicability of the gained findings.