The Proliferation and Countering of Hate and Countering Narratives on Social Media
Recent years have seen a surge in hateful speech on social media platforms, with many individuals shaping their perceptions of distant groups through such content. This project seeks to uncover factors and mechanisms that facilitate the spread of hate speech on X (formerly Twitter), and how such hateful messages are countered in online conversations. Focusing on hate speech against Jews and Muslims, we introduce a new annotation scheme, categorizing hate speech into thematic and rhetorical dimensions, as well as a new annotation scheme for the classification of speech that counters the hate (“counter hate speech”). Thematic categories outline different aspects of hate/counter speech, while rhetorical features analyze how the annotated messages are communicated. We are working to train a supervised prediction model to automatically identify hate and counter hate speech as defined in our project, which we will then use for exploring the dynamics of the two types of speech within conversation threads on X.
This study is part of a collaborative project with a team of researchers from the University of Toronto. It is funded by the University of Toronto - Hebrew University of Jerusalem Research and Training Alliance, and partly by the Vidal Sassoon International Centre for the Study of Antisemitism.