Seminal
The magazine and atlas of EU-funded researchA new publication from the web and AI team at Scott Charnley

Atlas / United Kingdom

Marie Skłodowska-Curie Action · Running

GENPOLAR Uncovering the Generative Mechanisms of Hate Speech and Opinion Polarization in Online Communities

Coordinator
Brunel University London, United Kingdom
Runs
1 January 2027 to 31 December 2030
Status
Running
EU funding
€346k
Total cost
—
Partners
5 organisations in 5 countries
Type
Marie Skłodowska-Curie Action
Grant agreement
101338810

What the project is doing

In the project's own words, from its summary on CORDIS.

Hate speech and opinion polarization threaten democratic resilience, social cohesion, and citizen well-being across the globe. While computational methods have advanced the detection of toxic speech online, the underlying mechanisms that drive polarization and escalate hate remain poorly understood. GENPOLAR addresses this gap by applying a critical realist process perspective to uncover the generative mechanisms of polarization and hate speech in diverse online contexts.

The project brings together a comparative partnership spanning Europe and Asia, organised around four national case studies. Brunel University London will investigate the effectiveness of Reddit's large-scale community bans (""the Great Ban"") in reshaping toxic user behaviour and mitigating hate speech. Università degli Studi di Salerno will examine migration and populist debate on Twitter/X and in the comment sections of major Italian news portals.

Insan Haklari Okulu Dernegi will study immigration discourse in Türkiye across online forums, social media and national news portals, contributing a human rights perspective on digital media. Yonsei University (South Korea) will analyse gender-based conflicts that have become flashpoints for extreme polarization in digital spaces. Budapesti Corvinus Egyetem contributes process mining and computational expertise to cross-case comparison.

Together, these case studies provide a cross-cultural basis to identify similarities and differences in the mechanisms that fuel polarization. GENPOLAR integrates computational social science methods (NLP, stance detection, network analysis, clustering) with qualitative inquiry to move beyond surface-level correlations. By developing process theories of polarization grounded in critical realism, the project will explain how interventions (such as community bans or moderation) succeed or fail in transforming online behaviour.

Through structured staff exchanges, interdisciplinary training, and comparative empirical work, GENPOLAR will build a lasting research network that bridges computational methods, social theory, and policy relevance. The project will generate actionable insights for scholars, platform managers, and policymakers seeking to counter online hate and polarization, while equipping a new generation of researchers with the skills to tackle one of the most urgent challenges.""

Who is involved

OrganisationCityCountryRole
Brunel University LondonUxbridgeUnited KingdomCoordinator
Budapesti Corvinus EgyetemBudapestHungaryAssociatedpartner
Universita Degli Studi Di SalernoFisciano SaItalyParticipant
Yonsei UniversitySeoulSouth KoreaAssociatedpartner
Insan Haklari Okulu DernegiÇankayaTürkiyeParticipant

Source: CORDIS, European Commission. Data of 5 Nov 2026, reused under the Commission's reuse policy. View on CORDIS · Grant DOI