Online Library

ISIS and other extremist groups, as well as their online supporters, have continued to exploit and misuse Google’s platforms to disseminate propaganda material, despite the company having repeatedly announced increased measures to combat online extremism.1 On July 21, 2017, Google announced the launch of one such measure––its Redirect Method Pilot Program. The program is intended to target individuals searching for ISIS-related content on YouTube and direct them to counter-narrative videos, which try to undermine the messaging of extremist
groups.2 The Counter Extremism Project (CEP) monitors and tracks ISIS and other terrorist organizations’ material on YouTube. Between August 2 and August 3, 2018, CEP reviewed a total of 649 YouTube videos for extremist and counter-narrative content. The result of CEP’s searches highlights the extent of the enduring problem of terrorist content on YouTube and undermines claims touting the efficacy of the company’s efforts to combat online extremism.

This article is based on a case study of the online media practices of the militant neo-Nazi organization the Nordic Resistance Movement, currently the biggest and most active extreme-right actor in Scandinavia. I trace a recent turn to humor, irony, and ambiguity in their online communication and the increasing adaptation of stylistic strategies and visual aesthetics of the Alt-Right inspired by online communities such as 4chan, 8chan, Reddit, and Imgur. Drawing on a visual content analysis of memes (N = 634) created and circulated by the organization, the analysis explores the place of humor, irony, and ambiguity across these cultural expressions of neo-Nazism and how ideas, symbols, and layers of meaning travel back and forth between neo-Nazi and Alt-right groups within Sweden today.

VOX-Pol Blog post.

VOX-Pol Blog Post.

Internet memes are increasingly used to sway and manipulate public opinion. This prompts the need to study their propagation, evolution, and influence across the Web. In this paper, we detect and measure the propagation of memes across multiple Web communities, using a processing pipeline based on perceptual hashing and clustering techniques, and a dataset of 160M images from 2.6B posts gathered from Twitter, Reddit, 4chan’s Politically Incorrect board (/pol/), and Gab, over the course of 13 months. We group the images posted on fringe Web communities (/pol/, Gab, and The_Donald subreddit) into clusters, annotate them using meme metadata obtained from Know Your Meme, and also map images from mainstream communities (Twitter and Reddit) to the clusters. Our analysis provides an assessment of the popularity and diversity of memes in the context of each community, showing, e.g., that racist memes are extremely common in fringe Web communities. We also find a substantial number of politics-related memes on both mainstream and fringe Web communities, supporting media reports that memes might be used to enhance or harm politicians. Finally, we use Hawkes processes to model the interplay between Web communities and quantify their reciprocal influence, finding that /pol/ substantially influences the meme ecosystem with the number of memes it produces, while td has a higher success rate in pushing them to other communities.

From its start, the so-called Islamic State of Iraq and the Levant (ISIL/ISIS) has been successfully exploiting social media networks, most notoriously Twitter, to promote its propaganda and recruit new members, resulting in thousands of social media users adopting pro-ISIS stance every year. Automatic identification of pro-ISIS users on social media has, thus, become the centre of interest for various governmental and research organisations. In this paper we propose a semantic-based approach for radicalisation detection on Twitter. Unlike most previous works, which mainly rely on the lexical and contextual representation of the content published by Twitter users, our approach extracts and makes use of the underlying semantics of words exhibited by these users to identify their pro/anti-ISIS stances. Our results show that classifiers trained from words’ semantics outperform those trained from lexical and network features by 2% on average F1-measure.