Detecting Online Extremist Networks Using Open Source Social Media: A Content-Centric Approach

This research note summarizes experimental research conducted by The SecDev Group in 2013, as part of a Public Safety Canada, Kanishka-funded project looking at social media analytics and the prevention of violent extremism. These experiments explored different ways in which content — keywords and in-group phrases, Facebook Interest Pages — could be used to surface online violent extremist networks, with case studies focused on the White Supremacist movement. Findings suggest that it is possible to identify and operationalize content-based hallmarks of violent extremist groups/online networks. Facebook Graph Search can also be effective in detecting members of communities at risk of radicalization toward violent extremism; however, the analysis and monitoring of FB Pages – even the public pages of members of extremist groups – could contravene the public’s evolving normative expectations around online privacy, even if framed as an ‘early warning and intervention’ activity.