Machine Learning Approaches for Extremism Detection: Challenges, Pitfall and Opportunities

Social networks, such as Facebook, Twitter, and YouTube, are free for everyone, including extremist groups. These extremist groups may use social media to spread radical ideas, disseminate propaganda, and recruit young people to promote their cause. Therefore, identifying and eradicating extremist content is among the top challenges for technology companies, counter-terrorism agencies and regulatory authorities. In this study, we have reviewed extremism detection in social media literature. In particular, a systematic literature review is conducted and a synthesis of the literature is performed to examine the strengths and weaknesses of the existing studies on extremism detection. Furthermore, the challenges are identified and opportunities for future research are discussed. We contend that the results of the study can be used by researchers for a better comprehension of the landscape of online extremism extension, as well as to open new horizons for the research community.