Online Library

VOX-Pol Blog post.

This study compares the narratives and language of QAnon groups in the encrypted messaging apps Telegram and Discord to those observed in the manifestos of terrorists. Drawing on our systematic linguistic analysis of fifteen terrorist manifestos that were published in the past decade, we developed a coding scheme which traces the narratives and linguistic markers that occur in the written communication of perpetrators of political violence. In this pilot study we apply our new coding scheme to QAnon content to assess the scale and nature of violence-associated narratives within the movement. Based on 200,000 messages that we collected from the online QAnon group “Great Awakening Community” on the gaming chat application Discord, we quantitatively examine to what degree they carry the trademarks of violent terrorist manifestos that are not found in non-violent texts. We then compared the results for the Great Awakening Community to content from both a non-violent and a violent-terrorist control group. To complement our computational assessment of QAnon narrative and linguistic patterns we share ethnographic observations from ten QAnon Telegram and Discord groups with English, German, and French speaking audiences. Past research has found that identity fusion in combination with a range of mediating and moderating variables is a strong predictor of violence in groups, and this is further supported by our terrorist manifesto analysis. Our study of QAnon messages found a high prevalence of linguistic identity fusion indicators along with external threat narratives, violence-condoning group norms as well as demonizing, dehumanizing, and derogatory vocabulary applied to the out-group, especially when compared to the non-violent control group. The aim of this piece of research is twofold: (i.) It seeks to evaluate the national security threat posed by the QAnon movement, and (ii.) it aims to provide a test of a novel linguistic toolkit aimed at helping to assess the risk of violence in online communication channels.

This paper examines in detail the role of the Internet in individual radicalization processes of eight German former right-wing extremists. Applying Grounded Theory methodology the qualitative interviews were analyzed in several coding and re-coding phases. The findings are integrated into the existing literature afterwards. Besides well known factors, such as more effective communication, anonymity and better networking opportunities, this study found evidence that the Internet is a major driving factor to establish and foster the development of radical contrast societies (cf. Koehler, 2015) transmitting radical and violent ideologies and translating them into political activism. As a venue for information exchange, ideological development and training, the individual radicalization process was characteristically shaped or even made possible through the Internet. This paper also shows the high value of qualitative research regarding the topic in contrast to usually employed quantitative analysis of webpage content.

VOX-Pol Blog Post.

In 2020, OpenAI developed GPT-3, a neural language model that is capable of sophisticated natural language generation and completion of tasks like classification, question-answering, and summarization. While OpenAI has not opensourced the model’s code or pre-trained weights at the time of writing, it has built an API to experiment with the model’s capacity. The Center on Terrorism, Extremism, and Counterterrorism (CTEC) evaluated the GPT-3 API for the risk of weaponization by extremists who may attempt to use GPT-3 or hypothetical unregulated models to amplify their ideologies and recruit to their communities.

This paper is a multimodal critical discourse study of self-representation in ISIS’s e-magazine, Dabiq, employing Social Movement Theory and Van Leeuwen’s Socio-semantic Inventory. By analyzing the linguistic and non-linguistic features in the representation of social actors and actions in Dabiq, ISIS’s implied ideology at the macro level, which is to convince the prospective recruits and at the same time, to legitimize its brutality, is revealed from both textual and visual perspectives. The results showed an interplay of the contradictory representational choices, namely “lionization” and “victimization” that has enabled ISIS to create a powerful narrative. By portraying its so-called “knights” as undefeatable lions and glorifying their acts of violence, while, excluding its killed, injured or imprisoned agents both textually and visually, ISIS has attempted to provoke admiration among its current or potential followers. Significantly, they simultaneously depict an air of victimhood to further legitimize their act of terror and represent themselves as the godsent saviours of the victims, and confer a sense of security in their hearts. Victimization, however, has rarely been applied to the Mujahidin of the State who are fighting at forefronts, but rather to the State itself as a territory (Caliphate) and to the Muslims who are presumably plagued by the enemies’ transgression and injustice.