Can Generative AI Eliminate Speech Harms? A Study on Detection of Abusive and Hate Speech during the COVID-19 Pandemic
Chen-Shu Wang, Heng‐Li Yang, Boyi Li, Hong-Yan Chen · 2023
This study investigates the issue of abusive and hateful language arising from the COVID-19 pandemic, this research employs machine learning techniques to establish a system capable of detecting and rephrasing abusive and hateful language. To begin, a dataset and dictionary specific to abusive and hateful language in Chinese. Subsequently, a two-stage detection model is proposed, with the BERT model yielding the most optimal outcomes. The first stage attains an accuracy of 94.42%, while the second stage achieves an accuracy of 81.48%.