Multimodal large language models can make context-sensitive hate speech evaluations aligned with human judgement

2025/12/15 by Thomas Davidson · 2 citations
Computer Science · #Audit #Context (archaeology) #Hate Speech and Cyberbullying Detection #Judgement #Language model #Lexical choice #Moderation #Sentiment Analysis and Opinion Mining #Social media #Spam and Phishing Detection

paper · doi:10.1038/s41562-025-02360-w

published in Nature Human Behaviour 10(3), 514-530 (Nature Portfolio)

openalex created_date 2025/12/15 · openalex publication_date 2025/12/15 · openalex updated_date 2026/07/28

Citations

• On the Dangers of Stochastic Parrots
• The social media discourse of engaged partisans is toxic even when politics are irrelevant
• AI generates covertly racist decisions about people based on their dialect

Cited by

• AI reasoning effort predicts human decision time in content moderation
• Assessing and mapping public visual perception across urban public space typologies based on geotagged social media images
• Dynamics of Cognitive Heterogeneity: Investigating Behavioral Biases in Multi-Stage Supply Chains with LLM-Based Simulation