Tag
A multi-stage agentic framework for generating, refining, and evaluating counter-narratives to combat hate speech and misinformation, with human validation showing improved persuasiveness and safety over expert-written counterspeech.