<![CDATA[This study presents a systematic literature review (SLR) on the automatic detection of cyberbullying across multiple media modalities, including text, images, and videos. The PRISMA framework was employed to guide the review process, focusing on articles published from 2020 to 2025 in the Scopus database. A total of 4272 records were initially identified, and after a structured screening process, 120 studies were included in the final synthesis. The results indicate an increasing scholarly interest in the topic, with a peak in 2025. Text-based detection remains the most prevalent approach, but there is a growing trend toward the integration of image and video analysis, particularly using multimodal and hybrid models. Commonly used techniques include Natural Language Processing (NLP), Convolutional Neural Networks (CNN), Support Vector Machines (SVM), and advanced models such as BERT and LSTM. The review also identifies gaps in real-time detection capabilities and the limited use of explainable AI. This study contributes to a deeper understanding of current methods, trends, and future research directions for cyberbullying detection systems]]>
Copyrights © 2026
