A Survey on Combating Hate Speech through Detection and Prevention in English (ACL Anthology)

The scientific study of hate speech from a computer science viewpoint has garnered prominence in recent years. Hate speech can be textual or multimodal information, such as memes, GIFs, audio, or movies. With annotated corpora and shared resources, hate speech detection and prevention are mostly regarded as supervised tasks. This survey thoroughly looks at the efforts made to prevent hatred in English, and social networks are applying contemporary AI techniques to counteract hate speech. It explores the effectiveness and drawbacks of cutting-edge techniques for counternarrative production, explainable AI, unimodal and multimodal hate detection, and style transfer-based hate speech prevention. In contrast to previous surveys, this report provides a coherent description of hate-fighting strategies.

https://aclanthology.org/2024.icon-1.57