Nature Human Behaviour

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

2025-12-15

Full text

DOI https://doi.org/10.1038/s41562-025-02360-w