Edinburgh Napier University

  The freedom of speech in online spaces has substantially promoted engagement on social media platforms, where cyberbullying has emerged as a significant consequence. While extensive research has been conducted on cyberbullying detection in English, efforts in the Arabic language remain limited. To address this gap, the current study provides a comprehensive, state-of-the-art review of datasets and methodologies specifically focused on Arabic cyberbullying detection. It systematically reviews different relevant studies from six academic databases, examining their methodologies, dataset characteristics, and performance in terms of classification accuracy and limitations. The paper critically evaluates existing Arabic cyberbullying datasets according to criteria such as dataset size, dialectal diversity, annotation processes, and accessibility. Additionally, this review identifies critical limitations, including dataset scarcity, dialectal imbalance, annotation subjectivity, and methodological constraints. By synthesizing current knowledge, identifying research gaps, and suggesting future directions, this review supports the development of more robust, effective, and linguistically inclusive analytical methods. Ultimately, this work contributes significantly to natural language processing research and advances the creation of safer online environments for Arabic-speaking users.