Detecting Defamation Words from Social Media Comments

The pervasiveness of social media has facilitated unprecedented communication and expression, but also fueled a surge in defamation cases. In Thailand, defamation carries significant legal weight, posing challenges for identifying and prosecuting perpetrators due to online anonymity and rapid content dissemination. This research proposed a novel machine learning approach to detect defamatory words in Thai social media comments. We utilized the state-of-the-art BERT model and the PyThaiNLP library, alongside a unique dataset of labeled comments from the Criminal Court of Thailand. This robust framework aimed to effectively identify potentially defamatory content, contributing to a safer and more responsible online environment.