Online Library

VOX-Pol Blog post.

With the ever-increasing cases of hate spread on social media platforms, it is critical to design abuse detection mechanisms to pro-actively avoid and control such incidents. While there exist methods for hate speech detection, they stereotype words and hence suffer from inherently biased training. Bias removal has been traditionally studied for structured datasets, but we aim at bias mitigation from unstructured text data. In this paper, we make two important contributions. First, we systematically design methods to quantify the bias for any model and propose algorithms for identifying the set of words which the model stereotypes. Second, we propose novel methods leveraging knowledge-based generalizations for bias-free learning. Knowledge-based generalization provides an effective way to encode knowledge because the abstraction they provide not only generalizes content but also facilitates retraction of information from the hate speech detection classifier, thereby reducing the imbalance. We experiment with multiple knowledge generalization policies and analyze their effect on general performance and in mitigating bias. Our experiments with two real-world datasets, a Wikipedia Talk Pages dataset (WikiDetox) of size ~ 96k and a Twitter dataset of size ~ 24k, show that the use of knowledge-based generalizations results in better performance by forcing the classifier to learn from generalized content. Our methods utilize existing knowledge-bases and can easily be extended to other tasks.

This report describes how terrorist, violent extremist and organized criminal groups are trying to take advantage of the Coronavirus disease (COVID-19) pandemic to expand their activities and jeopardize the efficacy and credibility of response measures by governments.

Although the subject of extreme right virtual community formation is often discussed, an online ‘sense of community’ among right-wing extremists has not been systematically analysed. It is argued that to study this phenomenon and to understand its backgrounds and function, the offline and online experiences and actions of those involved need to be taken into account. For this purpose, qualitative data has been collected on the web forum ‘Stormfront’, supplemented by extensive online interviews with eleven of its members. It is demonstrated that those experiencing stigmatization in offline social life regard the forum as a virtual community that functions as an online refuge, whereas those who – because of special circumstances – do not experience offline stigmatization do not display an online sense of community. It is concluded that offline stigmatization underlies virtual community formation by Dutch right-wing extremists. Because this mechanism may have broader significance, additional hypotheses for future research are formulated.

There has been a notable increase in support for far-right ideologies across the West. The seriousness of this threat has been acknowledged by the UK government which has banned certain far-right groups using terrorism legislation. While criminological theories have been useful in explaining general criminality, they have been under-utilised in explaining extremism and terrorism. Agnew’s General Strain Theory, which hypothesises that negative life events increase the chance of a turn to criminality, is explored in this article alongside Control Theory. Based on a survey (N 1,138) conducted on Facebook in late 2019, we explore how strain and resilience based on participants’ gender, economic situations, life events and their use of the internet impacts individuals’ far-right extremist attitudes and behaviours. We use regression analysis to investigate the impact that strain and resilience, individuals’ gender, economic situations, individual life experiences, and their use of the internet have on their propensity to associate, engage, and support far-right ideologies and linked violence. While strain is not found to be significant, resilience, gender and the use of the internet are.

This study examines the Pakistan military’s strategic use of social media in encouraging and sustaining public support for the ongoing war against terrorism in Pakistan. Its findings indicate that the military used significantly different types of strategic frames in response to a fast-changing, evolving security situation in the country. Framing was used strategically to facilitate public–military and people-to-people engagements. Motivational frames were the most dominant forms of communication used to generate dialogue between the military and the public in the war against terrorism and to enhance public participation in it. This study also indicates that different frames used by Pakistan’s military on social media significantly mediated military engagement with different segments of society during the critical phases of Pakistan’s ongoing war against terrorism.