Celebrity messages reduce online hate and limit its spread

2026-01-07

arXiv.org

AI Summary

This study addresses the rapid spread of online hate speech and the lack of scalable preventive interventions by conducting a 20-week randomized controlled trial on X (formerly Twitter) involving 73,136 Nigerian users previously engaged in ethnic hate speech. Participants received prosocial video messages recorded by local celebrities. The first large-scale evaluation of such an intervention demonstrates a significant reduction in both posting and sharing of hate content: hate-related posts declined by 2.5%–5.5% during the intervention period, with 75% of the effect persisting for four months post-intervention. Users with high exposure exhibited over a 50% reduction in resharing, and evidence of spillover effects within social networks was observed. Integrating behavioral nudges, social network analysis, and scalable digital delivery, this work provides empirical support for platform-level governance strategies against online hate.

Technology Category

• Application Domains: Humanities & Computational Social Science
• Natural Language Processing: Safety and Robustness
• Machine Learning: Adversarial Learning & Robustness

Application Category

• Responsible Web: Mitigating online harms, including harassment, spam, and fake reviews
• Economics, Online Markets and Human Computation: The sharing economy
• Social Networks and Social Media: Media and governance, opinion dynamics, filter bubbles, polarization

Abstract

Online hate spreads rapidly, yet little is known about whether preventive and scalable strategies can curb it. We conducted the largest randomized controlled trial of hate speech prevention to date: a 20-week messaging campaign on X in Nigeria targeting ethnic hate. 73,136 users who had previously engaged with hate speech were randomly assigned to receive prosocial video messages from Nigerian celebrities. The campaign reduced hate content by 2.5% to 5.5% during treatment, with about 75% of the reduction persisting over the following four months. Reaching a larger share of a user's audience reduced amplification of that user's hate posts among both treated and untreated users, cutting hate reposts by over 50% for the most exposed accounts. Scalable messaging can limit online hate without removing content.

Problem

• online hate
• hate speech prevention
• scalable intervention
• social media
• content amplification

Innovation

• scalable intervention
• randomized controlled trial
• online hate speech
• prosocial messaging
• celebrity influence