Because of the increasing spreading of hate speech, in particular in social media, the detection of hateful contents is becoming a crucial task. For addressing this task NLP and sentiment analysis techniques are developed which are based on machine learning and annotated linguistic corpora. The evaluation of hate speech detection systems is among the activities proposed within the context of the more recent evaluation campaigns for computational linguistics tools.
We will describe our experience related to projects for countering and preventing hate speech ("Hate Speech & Social Media" and "Immigrants, Hate Speech and Prejudice in Social Media", http://hatespeech.di.unito.it/) and in the organization of evaluation campaign tasks about hate speech against different targets and in different languages, i.e. in Evalita 2018 for Italian (Hate Speech Detection task - HaSpeeDe) and SemEval 2019 for Spanish and English (task 5 - hatEval - Multilingual detection of hate speech against immigrants and women in Twitter).