LAUSR

In order to serve as an antidote to extremist messages, counter-messages (CM) are placed in the same online environment as extremist content. Often, they are even tagged with similar keywords. Given that automated algorithms may define putative relationships between videos based on mutual topics, CM can appear directly linked to extremist content. This poses severe challenges for prevention programs using CM. This study investigates the extent to which algorithms influence the interrelatedness of counter and extremist messages. By means of two exemplary information network analyses based on YouTube videos of two CM campaigns, we demonstrate that CM are closely-or even directly-connected to extremist content. The results hint at the problematic role of algorithms for prevention campaigns.