This paper conducts comprehensive examination of machine learning and deep learning methodologies applied in cyberbullying detection. The examination is conducted across diverse social media platforms by evaluating the efficacy of various algorithms such as support vector machines and convolutional neural networks. Additionally, the paper examine the psychological ramifications of cyberbullying and advocates for the integration of multimedia data sources to bolster detection capabilities with the ultimate aim of forging more effective solutions to combat cyberbullying in the digital age.