counterspeech

This paper studies the use of large language models to assist expert counterspeech writing when hate speech and misinformation co-occur, testing knowledge-driven strategies with human evaluation. The mixed strategy combining fact-checkers' and NGOs' guidelines proved most effective.