Online Library

The scientific study of hate speech, from a computer science point of view, is recent. This survey organizes and describes the current state of the field, providing a structured overview of previous approaches, including core algorithms, methods, and main features used. This work also discusses the complexity of the concept of hate speech, defined in many platforms and contexts, and provides a unifying definition. This area has an unquestionable potential for societal impact, particularly in online communities and digital media platforms. The development and systematization of shared resources, such as guidelines, annotated datasets in multiple languages, and algorithms, is a crucial step in advancing the automatic detection of hate speech.

Hateful and antagonistic content published and propagated via the World Wide Web has the potential to cause harm and suffering on an individual basis, and lead to social tension and disorder beyond cyber space. Despite new legislation aimed at prosecuting those who misuse new forms of communication to post threatening, harassing, or grossly offensive language – or cyber hate – and the fact large social media companies have committed to protecting their users from harm, it goes largely unpunished due to difficulties in policing online public spaces. To support the automatic detection of cyber hate online, specifically on Twitter, we build multiple individual models to classify cyber hate for a range of protected characteristics including race, disability and sexual orientation. We use text parsing to extract typed dependencies, which represent syntactic and grammatical relationships between words, and are shown to capture ‘othering’ language – consistently improving machine classification for different types of cyber hate beyond the use of a Bag of Words and known hateful terms. Furthermore, we build a data-driven blended model of cyber hate to improve classification where more than one protected characteristic may be attacked (e.g. race and sexual orientation), contributing to the nascent study of intersectionality in hate crime.

This article applies instruments of social network analysis to a study of communication networks within the Italian and German extremist right. Web links between organizational websites are used as a proxy. Indeed, extremist groups increasingly use and abuse the Internet for their propaganda and their recruitment, and also for their internal communication. The analysis includes both political parties and non-party organizations, even violent groups. In a macro-, micro-, and meso-analysis, the various specificities of the two national political sectors are demonstrated and linked to the offline reality. The Italian network appears to be very fragmented, highly diversified, and difficult to be coordinated (‘policephalous network’), whereas the German network is denser and much more concentrated on a few central actors (‘star structure’). These differences are mainly due to political opportunity structures in the two countries. Additionally, whereas the Italian network structure allows for the construction of a typology of sub-groups of organizations, the German communicative structure seems to be more erratic and less coordinated. The article also highlights the function of websites which are not related to any specific group. Indeed, these are of special importance for the far right as a political arena which is usually banned from the dominant societal discourses (if not even legally forbidden). Considering this, new modes of communication can be of greater use for extremist groups than for more traditional political actors.

A key challenge for automatic hate-speech detection on social media is the separation of hate speech from other instances of offensive language. Lexical detection methods tend to have low precision because they classify all messages containing particular terms as hate speech and previous work using supervised learning has failed to distinguish between the two categories. We used a crowd-sourced hate speech lexicon to collect tweets containing hate speech keywords. We use crowd-sourcing to label a sample of these tweets into three categories: those containing hate speech, only offensive language, and those with neither. We train a multi-class classifier to distinguish between these different categories. Close analysis of the predictions and the errors shows when we can
reliably separate hate speech from other offensive language and when this differentiation is more difficult. We find that racist and homophobic tweets are more likely to be classified as hate speech but that sexist tweets are generally classified as offensive. Tweets without explicit hate keywords are also more difficult to classify.

Race and racism persist online in ways that are both new and unique to the Internet, alongside vestiges of centuries-old forms that reverberate significantly both offline and on. As we mark 15 years into the field of Internet studies, it becomes necessary to assess what the extant research tells us about race and racism. This paper provides an analysis of the literature on race and racism in Internet studies in the broad areas of (1) race and the structure of the Internet, (2) race and racism matters in what we do online, and (3) race, social control and Internet law. Then, drawing on a range of theoretical perspectives, including Hall’s spectacle of the Other and DuBois’s view of white culture, the paper offers an analysis and critique of the field, in particular the use of racial formation theory. Finally, the paper points to the need for a critical understanding of whiteness in Internet studies.

Hate speech detection on Twitter is critical for applications like controversial event extraction, building AI chatterbots, content recommendation, and sentiment analysis. We define this task as being able to classify a tweet as racist, sexist or neither. The complexity of the natural language constructs makes this task very challenging. We perform extensive experiments with multiple deep learning architectures to learn semantic word embeddings to handle this complexity. Our experiments on a benchmark dataset of 16K annotated tweets show that such deep learning methods outperform state-of-the-art char/word n-gram methods by ~18 F1 points.