CATCH-ME if you RAG: a dataset of Contextually Annotated multi-Turn Counterspeech against Hate and Misinformation Exchanges

2026-06-18

AI Summary

This work addresses the gap in existing research, which often treats hate speech and misinformation in isolation and lacks high-quality, multilingual, fact-grounded multi-turn dialogue data. To bridge this gap, the authors introduce the first large-scale, expert-annotated multilingual multi-turn dialogue dataset that specifically targets scenarios where hate speech and misinformation intersect. Spanning five languages and seven marginalized groups, the dataset is anchored to external credible knowledge sources—such as fact-checking articles and NGO reports—and supports training retrieval-augmented generation (RAG) systems. It features document- and snippet-level knowledge citations, cross-lingual alignment, and fine-grained contextual annotations, enabling direct use for training and evaluating counter-speech generation models that are both persuasive and factually accurate.

Abstract

Online hate speech and misinformation frequently overlap, yet NLP research has mainly treated them in isolation. While LLMs represent a scalable solution for assisting humans in the generation of counterspeech for both threats, zero-shot models frequently generate repetitive and vague responses, underscoring the need for high-quality examples to steer model generation. However, existing counterspeech datasets against the overlap of hate and misinformation are scarce and limited to single-turn English dialogues, while real-life interactions span across multiple turns and languages. To bridge this gap, we introduce the first large-scale, expert-curated, multilingual dataset of dialogues tackling the intersection of hate and misinformation. To ensure factual grounding, the dialogues are also anchored in verified external knowledge (i.e., fact-checking articles and NGO reports) and include document- and chunk-level span annotations, making it directly applicable for RAG systems. Covering five languages and targeting hate directed at seven marginalized groups, this novel resource enables the training and evaluation of more persuasive, factually grounded counterspeech models.

Problem

• hate speech
• misinformation
• counterspeech
• multilingual
• multi-turn dialogues

Innovation

• counterspeech
• retrieval-augmented generation (RAG)
• multilingual dialogue
• hate and misinformation intersection
• fact-grounded annotation