Hateful Content

Hateful content online, encompassing hate speech and offensive language, is a significant research area aiming to develop automated detection and mitigation strategies. Current research focuses on improving detection accuracy using various machine learning models, including transformer-based architectures (like BERT and its variants) and graph neural networks, often incorporating multimodal data (text, images, audio) and contextual information from user networks and online discussions. These advancements are crucial for creating safer online environments and informing effective content moderation strategies, while also highlighting the challenges of cross-lingual generalization, bias in datasets, and the need for explainable AI in this sensitive domain.