Skip to content


The methodology of this proposal includes a set of activities developed across 5 interconnected work packages with the participation of the three members of the consortium (USAL, AUTH and UMIL). PHARM will first identify hate speech contents and understand how they can predict real hate crimes, and later it will use data-driven news pieces (with interactive journalistic data information created by journalists) and first-person testimonies (with dramatic, fictional and narrative structures) to counter hate speech.

PHARM will implement a conceptual and methodological common framework for large-scale analysis and detection of hate speech against refugees and migrants.

Based on an under-developing technical prototype, we will use Natural Language Processing and ML techniques to detect hate speech and store hateful contents, creating the most complete and articulated database of online hate speech in southern Europe, useful to pursue illegal hate speech, model its relationship with hate crime and create counter-narratives.

Monitoring hate speech

PHARM will then model hate crime based on the descriptive features of hate speech against refugees and migrants in order to predict future hate crime episodes.  We will create a database of hate crimes and will use cutting-edge ML algorithms to model the occurrence of physical/verbal aggressions against migrants and refugees.

Modelling hate speech and hate crime

PHARM will also create and disseminate counter-narrative fictional stories using narrative persuasion, and will evaluate its effects in countering the production of hate speech.