Online Library

Mainstream risk assessment frameworks such as TRAP-18, ERG22+, VERA-2R, and RADAR largely use Structured Professional Judgement to map individuals against four critical factors; ideology, affiliation, grievance, and moral emotions. However, the growing use of online communication platforms by extremists presents a series of opportunities to complement or extend existing risk assessment frameworks. Here, we examine linguistic markers of morality and emotion in ideologically diverse online discussion groups and discuss their relevance to extant risk assessment frameworks. Specifically, we draw on social media data from the Reddit platform collected across a range of community topics. Nine hundred and eighty-eight threads containing 272,298 individual comments were processed before constructing high-order models of moral emotions. Emotional and moral linguistic content was then derived from these comments. We then conducted comparisons of linguistic content between mainstream left and right political discourse, anti-Muslim (far-right), Men’s Rights (Incel-like), and a nonviolent apolitical control group. Results show that a combination of individualising moral communication and high emotionality separate far-right and Incel-like groups from mainstream political discourse and provide an early warning opportunity.

Much qualitative research has drawn an association between social media and domestic terrorism, with the studies reaching different conclusions. However, few empirical studies have evaluated whether the surge in social media participation affects domestic terrorist events. Controlling for common explanations in the literature, we conduct a cross-national, time-series analysis of up to 151 countries from 2000 to 2019 to assess the impact of social media penetration on domestic terrorism. We find that greater social media penetration increases the likelihood of domestic terrorism in countries as it supports extremists’ ability to recruit, mobilize, and train terrorists. Using mediation analysis, we also find that greater social media penetration amplifies online and political polarization, increasing the likelihood of domestic terrorism events. Our work indicates the possible mechanisms linking social media and domestic terrorism and the need to develop and apply appropriate counterterrorism strategies to mitigate terrorist operations.

This report compares two large random samples of Twitter accounts that tweet in English: one taken just before Elon Musk acquired Twitter in October 2022, and one taken three months later, in January 2023. It also examines several related datasets collected during the period following the acquisition, a period in which, the study found, new accounts were created at a record-breaking pace. Some extremist and conspiracy networks created accounts faster than the baseline rate, probably because changes to Twitter’s trust and safety policies had been announced. In the context of these policy announcements, the study examines some reinstated accounts, with mixed results. Despite the loosening of several content policies, accounts that automated the sending of tweets (‘bots’) saw activity drop sharply during the period of the study, with many bot accounts being suspended or deactivated, while others voluntarily curtailed their activity in light of the API changes announced. Deactivated accounts were dominated by sex- related content and apparent financial spam or scams, often coupled with automated tweeting.

It is well established that jihadist groups and their supporters post URLs on online platforms to outlink to items of propaganda stored on other platforms. Industry initiatives – such as Global Internet Forum to Counter Terrorism’s inclusion of URLs in its hash-sharing database, and Tech Against Terrorism’s Terrorist Content Analytics Platform – have sought to counter this practice. These measures, together with new regulatory regimes (e.g. the EU’s Terrorist Content Online Regulation) and the growing use of decentralised services, raise the question whether jihadist groups’ propaganda dissemination strategies are perhaps being forced to evolve. This report considers whether there is evidence of such an evolution, by examining the means that three jihadist groups (Islamic State (IS), Al-Qaeda (AQ) and Al-Shabaab) used to disseminate their propaganda during a two-month period in early 2023.

Over the past decade, two major phenomena have developed in the digital realm. On the one hand, extremism has grown massively on the Internet, with sprawling online ecosystems hosting a wide range of radical subcultures and communities associated with both ‘stochastic terrorism’ and the ‘mainstreaming of extremism’. On the other hand, Artificial Intelligence (AI) has undergone exponential improvement: from ChatGPT to video deepfakes, from autonomous vehicles to face-recognition CCTV systems, an array of AI technologies has abruptly entered our everyday lives. This report examines ‘AI extremism’, the toxic encounter of these two evolutions – each worrying in its own right. Like past technological progress, AI will indeed be – in fact already is – used in various ways to bolster extremist agendas. Identifying the many opportunities for action that come with a range of AI models, and linking them with different types of extremist actors, we offer a clear overview of the numerous facets of AI extremism. Building on the nascent academic and government literature on the issue as well as on our own empirical and theoretical work, we provide new typologies and concepts to help us organize our understanding of AI extremism, systematically chart its instantiations, and highlight thinking points for stakeholders in countering violent extremism.

VOX-Pol Blog post.