A study of Hate Speech in Social Media during the COVID-19 outbreak

Viviana Cotik, Natalia Debandi, Franco M. Luque, Paula Miguel, Agustín Moro, Juan Manuel Pérez, Pablo Serrati, Joaquin Zajac, Demián Zayat

Keywords: hate speech, COVID-19, social media, machine learning

TL;DR: Interdisciplinary work in progress on building a corpus for hate speech in Spanish and studying hate speech phenomena in the pandemic context.

Abstract:

In pandemic situations, hate speech propagates in social media, new forms of stigmatization arise and new groups are targeted with this kind of speech.

In this short article, we present work in progress on the study of hate speech in Spanish tweets related to newspaper articles about the COVID-19 pandemic.

We cover two main aspects: The construction of a new corpus annotated for hate speech in Spanish tweets, and the analysis of the collected data in order to answer questions from the social field, aided by modern computational tools.

Definitions and progress are presented in both aspects. For the corpus, we introduce the data collection process, the annotation schema and criteria, and the data statement. For the analysis, we present our goals and its associated questions. We also describe the definition and training of a hate speech classifier, and present preliminary results using it.