2025/03/03 by Javier J. Amores, Carlos Arcila Calderón · 1 voice · 3 citations
Computer Science · Social Sciences · #Hate Speech and Cyberbullying Detection #Populism, Right-Wing Movements
paper · pdf · doi:10.15581/003.38.1.013
openalex publication_date 2025/03/03 · openalex created_date 2025/03/04 · openalex updated_date 2026/07/15
One of the greatest challenges facing democratic societies today is hate speech, that spreads massively and uncontrollably through social media. particularly racist and xenophobic speech, the category of discrimination in which most hate crimes are recorded annually in Southern Europe. In this context, many studies have already focused on analysing hate on X (formerly Twitter), but few have studied other platforms or specifically focused on messages directed at migrants and refugees, or in languages other than English. The present work aims to analyse, using computational methods, anti-immigration hate speech spread through Twitter messages and YouTube comments, in the contexts and languages of the main Southern European countries: Spain, Italy, and Greece. Specifically, after conducting a manual classification of messages about migrants and/or refugees on both platforms and in the three Mediterranean countries, the most frequent words were analysed, and topic modelling was applied to the messages classified as racist and/or xenophobic. In general terms, the underlying topics in these messages mostly identify immigrant and/or refugee groups as a realistic and/or symbolic threat to the receiving countries, which is directly related to the findings of previous works where media and audience frames of negative representations of migration were studied. Furthermore, in all three countries, there appears to be a predominance of hate directed at Arab and Muslim communities, which stands out in comparison to the proportion of hate aimed at other outgroups.