A multi-agent method improves responses to hate speech online
A Multi-Stage Agentic Framework for Effective Counter-Narrative Generation and Refinement
Computation and LanguageComputers and SocietySocial and Information Networks
Summary
Hate speech and misinformation spread quickly on social media, and trying to stop them directly can make things worse. The authors show a new way to create and improve messages that respond to hateful or false claims using multiple AI agents working together. For example, they focus on pro-Russian misinformation about the war in Ukraine and find that some styles, like repeating ideas with emotional tones, make responses more effective. Their approach outperforms basic AI responses and matches expert-written replies in safety and impact. This method could help social platforms or organizations craft better counter-messages against harmful content.
What this means in practice
- •For social media moderators: Generate and improve counter-messages tailored to specific harmful narratives to reduce polarization and enhance user engagement.
- •For public relations teams: Develop targeted responses to misinformation campaigns that effectively counter false narratives while maintaining emotional appeal.
Authors
Carmel Kronfeld, Sharva Gogawale, Tetsuro Kobayashi, Irad Ben-Gal
Abstract
The rapid diffusion of hate speech and misinformation on social networks challenges democratic societies, since direct suppression efforts may deepen polarization, fuel public distrusts, and strengthen extremist narratives. LLM-driven counter-narratives (CNs) offer a promising way to reduce those risks, yet their effectiveness depends on rhetorical and stylistic choices that remain poorly understood. We present a multi-stage agent-based framework for generating, refining, and evaluating CNs, applied to pro-Russian hate and misinformation narratives on the war with Ukraine and adaptable to other domains. A pilot experiment with human evaluators identifies effective technique style pairings, such as repetition with emotional framing enhancing persuasiveness. Building on these insights, we introduce a multi-agent refinement process that iteratively improves CNs for persuasiveness, emotional engagement, and shareability. After human validation confirmed improvement, an automated safety analysis shows that our refined CNs match or improve on expert-written counterspeech. A simulated experiment then shows that they reduce the perceived strength of pro-Russian narratives and consistently outperform a vanilla LLM baseline, highlighting a pathway toward scalable, narrative-specific interventions against hate speech and misinformation. Code and data accompanying this work are publicly available at https://github.com/carmelkron/inlg2026-counter-narratives.