Countering Online INtolerance against especially vulnerable groups. Large-scale monitoring and narrative combat of online hatred in Spain (COIN)

The COIN project (Countering Online INtolerance) addresses online intolerance against particularly vulnerable groups by developing advanced tools to detect, track, and analyze hate speech online. This project, stemming from the PHARM program (Preventing Hate Against Refugees and Migrants) [https://pharmproject.usal.es/], utilizes artificial intelligence technologies, such as deep learning and supervised machine learning, to enhance the detection of hate speech based on various prejudices like racism, xenophobia, anti-Muslim sentiment, antisemitism, anti-Romani sentiment, and LGBTIQphobia, including cases where these prejudices intersect.


  1. Automated Tools Development (SO1): Create an automatic detector that improves the identification and monitoring of hate content. Use a broad and manually tagged corpus to train models adapted to each type of hate and its propagation on the web.
  2. In-depth Analysis of Hate Dissemination (SO2): This involves integrating computational techniques and traditional methods to generate detailed maps of hate speech, complemented by bibliometric analysis and data mining, to qualitatively understand the social impact of this phenomenon in Spain.
  3. Social Intervention Strategy (SO3): Implement a social intervention strategy based on narrative persuasion, designed to increase empathy, and reduce the effects of hateful content, leveraging insights gained in earlier phases of the project.


The project develops in two main phases:

  1. First Phase: Use of big data and AI techniques to understand hate speech from both quantitative and qualitative perspectives, using advanced tools for content access on social networks and analysis at the Castilla y León Supercomputing Center (Scayle).
  2. Second Phase: Application of the knowledge acquired to design and implement effective social interventions with audiovisual producers, using narrative audiovisual stories adapted to different audiences.

The COIN project not only represents a technical advancement in the automatic detection of online hate but is also a pioneer in applying an interdisciplinary approach that includes media psychology, machine learning, and data analysis. This approach is complemented by the participation of professionals from the audiovisual industry, enriching the development of adapted and effective social intervention strategies.

COIN integrates a gender perspective at all project stages to ensure equity and avoid biases in detecting hate speech. Additionally, it adheres to the principles of Responsible Research and Innovation (RRI), promoting gender equality, scientific education, and open access, ensuring an ethical and participatory approach in all project activities.

In collaboration with:

Project funded by the European Union. European Union’s Actions-H2020-MSCA-COFUND-2020-USAL4EXCELLENCE-PROOPI-663. Grant agreement: N. 101034371. USAL internal reference: 8925-8553.