Radicalisation of the Online Far-Right: Comparing Mainstream to Fringe Web Platforms

Far-right extremist violence has recently increased across the world, and the internet has become a vital tool for far-right extremists to radicalize, recruit, mobilize, and network. This has resulted in a sharp rise in online hate crimes, and consequently a number of far-right groups have recently been banned from Facebook and Twitter, the two largest mainstream social media sites. Following these bans, a number of these groups have migrated to more fringe internet platforms such as Gab and 4Chan. How the online dynamics of these groups differ between mainstream platforms and these more marginal sites, where regulation, moderation and counter-speech are less common, remains unclear. A key measure of these conversations is online hate speech, the negative outcomes of which include increased racism, prejudice, and even offline hate crime. Here we use novel automated methods to detect and monitor hate speech online to investigate how hate speech develops and how users within far-right communities online spread this material on different platforms. More specifically, we build a cross-platform hate speech detection system, and use it to measure the radicalisation profiles for users on mainstream social media platform Facebook and fringe platform Gab. We show that these platforms give rise to different radicalisation trajectories over time for users taking part in the conversations, demonstrating the importance of taking platform-specific approaches and opening up questions about how counter speech and counter extremism work could be best applied on these different platforms.