Online Library

In recent years, terrorism and radicalization have been a consistent issue that many countries have faced. The Islamic State of Iraq and Syria (ISIS) has been the most recent in a long trail of organizations that have sought to strike terror against the western world. However, ISIS is distinguished from other groups, like Al-Qaeda, in that ISIS supports a complex propaganda machine. Although ISIS is not the first organization to use the social media platform, they are the first to use it with such diversity. The two main channels that ISIS uses to spread their propaganda messages are through social media sites such as Twitter and through online journals such as the Dabiq. Recent research has attempted to determine how recruitment messages are being received and which messages trigger recruitment. It is the goal of this paper to determine which messages are salient, and the psychological constructs that support them. By coding messages for appeals to identity, need for cognitive closure, time pressure, and appeals to ideology, the researchers expect that the two main channels of ISIS propaganda differ in their messages. We hypothesize that Twitter messages will be targeted towards novice ISIS sympathizers, whereas the Dabiq will be focused on already-radicalized individuals who have moved past the introduction of the radical ideology.

With the popularity and ease of using social media platforms, users are able to, in varying capacities, connect with others in varying capacities. During 2015, there were approximately 305 million worldwide active monthly Twitter users. While Twitter has maintained implementation of their counter-extremism policies, supporters of ISIS have found ways to navigate around them and continue to use their platform as a means to connect with others. With 140 characters per Tweet, ISIS supporters are able to recruit and promote propaganda quickly with users around the world. By using a data set containing 16,841 Tweets from 104 ISIS supporters following the November 2015 attacks in Paris, content analysis will be conducted on the tweet itself to look for reoccurring themes and keywords. By understanding the keywords and reoccurring themes, military, law enforcement, and private sector counterterrorism units can better understand and implement policies and procedures relating to ISIS and the ways that they continuously navigate around the counter-extremism policies on Twitter.

This research examines the relationship between race and exposure to online hate material. The utilization of websites, weblogs, newsgroups, online games, radio broadcasts, online newsletters and a myriad of other online platforms has proliferated race-based hate groups in the US (Shafer 2002). According to the Southern Poverty Law Center (SPLC), the number of hate groups has been on the rise since the 1990s and continues to gain momentum with the advent of social media (Potok 2017). Exposure to separatist ideologies has propelled these radical rightwing groups into the mainstream by way of social media platforms, as they are “the most active producers of online hate material” (Costello, Hawdon, Ratliff, and Grantham 2016: pg. 313). That dissemination of radical rightwing ideologies, such as white supremacy, racial purity, and racial solidarity, exists is not enough in understanding what individuals are exposed to race-based hate ideologies in online platforms. Exposure is the key to understanding the growth of these race-based hate groups and ways of countering the efforts to disseminate radical rightwing ideologies due to its relationship to hate group emergence and persistence. More so, understanding how these groups target individuals and recruit through social networking sites can provide insight into exposure. Exposure to hate material aids groups in recruiting new members and victimizing potential targets. In the same manner, exposure to hate material is victimization of those who are exposed. In a sample collected by Costello et al. (2016a), of those exposed to hate material online nearly half centered on race. Thus, it is tantamount that research is conducted examining the role that race plays in determining who is exposed to hate material online, and how individuals react to hate material based on race. This dissertation will examine the importance of exposure to hate. Specifically, this dissertation will analyze survey data gathered from the Online Extremism Survey using logistic regression analysis and linear regression to understand exposure to hate material online and routine activity theory.

This study identifies and compares the methods of recruitment used by three prime jihadist organizations through their online magazines. The successful recruitment efforts and growth as a threat by the Islamic State of Iraq and Shām (ISIS), Al-Qaeda, and Al-Shabaab are attributed, in part, to the widespread popularity and accessibility of Dabiq and Rumiyah (published by ISIS), Inspire (published by Al-Qaeda), and Gaidi Mtaani (published by Al-Shabaab). In order to best examine the techniques of persuasion and propaganda to recruit new followers, the theoretical approach of compliance-gaining theory and methodological approach of narrative analysis are applied. The author analyzed a total of twelve magazine issues to compare how the four magazines use (1) imagery, (2) attrition, (3) intimidation, (4) propaganda, (5) spoiling, (6) outbidding, (7) incitement, and (8) recruitment. To validate each of these themes, the study applies an open-coding instrument to select each label based on specific dimensions. These findings reveal how the same purpose to achieve the Caliphate can be chosen on a similar path by these three different jihadist organizations. This path may take different turns at certain spots, but ultimately the road is based on the same historical context to justify the recruitment process. Overall, this study provides fresh descriptive insights on jihadist organizations’ recruitment methods to gain new followers toward the achievement of the Caliphate (i.e., the global Islamic state). The sample reveals that the online jihadist magazines integrate all analyzed themes to portray their messages. The findings show certain issues implement specific themes more than others. Across the sample, the use of persuasion justified and encouraged violence. Persuasion is found in prideful examples of successful terrorist attacks and aftermath scenes of destruction. The distinct difference between the four online jihadist magazines is the persuasion of recruitment.

This thesis asks why and how terrorist organizations use the Internet to achieve three strategic goals: 1) the dissemination of propaganda, 2) recruitment and 3) fundraising. It is immediately apparent that the Internet offers a number of advantages, including low cost, global reach, and anonymity. Nonetheless, terrorist organizations vary in their exploitation of these advantages according to their immediate objectives. To explain these variations, this thesis presents a comparative study of Al Qaeda, the Islamic State of Iraq and Syria (ISIS), and the Taliban. This study considers how different objectives render distinct narratives and thereby affect how those narratives leverage images and information in the dissemination of propaganda. Similarly, targets of recruitment vary according to the objectives of the respective organizations; this primarily affects their use of social media, including Facebook, YouTube, Twitter, and other applications. Fundraising methods also vary, from local means (extortion, narcotics, smuggling) to contemporary exploitation of digital platforms like PayPal and cell phone applications enabling anonymous donations. This thesis concludes that the sophistication of terrorist organizations online requires an equally sophisticated response that is as essential to the fight against violent extremism as kinetic operations.

Moderators of online discussion forums often struggle with controlling extremist comments on their platforms. To help provide an efficient and accurate tool to detect online toxicity, we apply word2vec’s Skip-Gram embedding vectors, Recurrent Neural Network models like Bidirectional Long Short-term Memory to tackle a toxic comment classification problem with a labeled dataset from Wikipedia Talk Page. We explore different pre-trained embedding vectors from larger corpora. We also assess the class imbalance issues associated with the dataset by employing sampling techniques and penalizing loss. Models we applied yield high overall accuracy with relatively low cost.