Due to the rapid use of AI on social media, algorithmic approaches are being explored for the detection, prevention, and mitigation of cyberbullying behaviors. However, Previous studies in this field have only attempted to expand on the emotional and psychological effects of cyberbullying, while very little attention has been paid to studies evaluating AI-based approaches for the prevention of cyberbullying on social media. This systematic literature review attempts to bridge this gap by synthesizing peer-reviewed studies, designs, and frameworks published between 2020 and 2025. These studies were obtained from different digital libraries such as Google Scholar, ScienceDirect, IEEE Xplore, Wiley Online Library, and the ACM Digital Library. The review examine the role of AI in addressing the threats to interpersonal interactions on social media. A total of 25 studies were included for synthesis in the narrative format and analyzed thematically. This literature review evaluates the current application of AI and examines the algorithmic evolution of digital safety from primitive keyword-based filters to improved AI frameworks capable of identifying nuanced patterns of cyberbullying. It also highlights both the possibilities and limitations of these technologies, particularly in relation to digital safety measures. Finally, after evaluating the current role of AI and its effectiveness, this literature review attempts to articulate the possible future directions of AI algorithms development to enhance their role in making social media a safer environment for online communication.
Glory Akwa (Thu,) studied this question.