Deep Learning-Driven AI System for Multimodal Detection and Classification of Harmful Content on Social Networks

This article addresses the significant issue of using deep learning techniques to detect and classify dangerous material in the framework of social networks. The ultimate goal of this project is to develop an artificial intelligence (AI) system that can identify unsuitable content across various media formats, including text, audio, and images. The system may identify potentially unsafe items using technologies such as Optical Character Recognition (OCR), Google Text-to-Speech (GTTS), and Natural Language Processing (NLP). Nowadays, this field of research must cope with problems including biased model outcomes and small datasets. The study comprehensively evaluates the efficacy of various strategies using standard measures, including accuracy, precision, recall, and F1measure. The findings indicate that, particularly with the Recurrent Neural Network (RNN) architecture, deep learning models are very successful. By facilitating the early detection and prevention of cyberbullying, our study contributes substantial new information towards the goal of creating safer and more inclusive societies, skillfully extracting data from social media platforms. Our models outperform prior research in identifying and classifying foul language thanks to our use of sophisticated preprocessing techniques and meticulous hyperparameter adjustment.