IMSyPP -Innovative Monitoring Systems and Prevention Policies of Online Hate Speech – is an EU REC AG project, ID 875263.

The IMSyPP project is tackling hate speech in a multidisciplinary fashion combining machine learning, computational social science and linguistic approaches to support a data-driven approach to hate speech regulation, prevention and awareness-raising.

The IMSyPP project addresses the following needs.

  • Automated detection and sustainable monitoring of hate speech is being addressed by near-real-time hate speech detection models tuned to language, culture and legislation, taking into account the context of the message.
  • Understanding of triggers and facilitators of hate speech spreading by identifying terms, entities, concepts that appear to trigger hate speech in online discussions.
  • Alternative- and counter-narratives emphasize the need to de-construct and weaken violent narratives.
  • Increase awareness and media literacy on all levels: the general audience, especially the youth, civil servants, educators, journalists, media players, on-line platforms, regulators and politicians.
  • Resources for science in form of high quality annotated corpora.

The IMSyPP project is developing models for automated hate speech detection and tracking four European languages (English, Italian, Slovenian and Dutch). By applying the models to a large dataset, we will study hate speech manifestations chronologically and geographically.

We will focus on different aspects:

  • Who is the target of hate speech,
  • Is there illegal hate speech,
  • The speed of spreading,
  • The audiences reached.

We will identify terms, entities and concepts to understand the triggers and facilitators of hate speech in online discussions. Facilitators and the role of instigators and trolls will be analysed. Furthermore, we will study the effect of journalistic techniques (newspaper articles, interviews, online infographics, videos, images …) on users’ reactions.

We are studying the impact of previous counter and alternatives in online exchanges. Based on the results, data-driven recommendations for counter/alternative narratives will be developed. Target groups for this objective will be identical to those of understanding triggers and facilitators of hate speech.

The IMSyPP project will facilitate further hate speech research by providing repositories of multi-lingual datasets, trained models in form of APIs and methodologies for the development of such models for other languages.