Multilingual AI system for detecting offensive content across text, audio, and visual media
Venkatesh Koreddi, Nalluri Manisha, Shaik Mohammad Kaif, Yeligeti Tejaswa Sai Kumar · Engineering Research Express · 2025
Abstract This paper aims to develop an AI system with advanced capabilities to detect offensive language across diverse platforms-covering text, audio (both live and recorded speech), and images (such as memes)—in multiple languages. Using technologies such as natural language processing (NLP), speech recognition (SR), and optical character recognition (OCR) to identify text within images, the system can already flag potentially harmful or inappropriate content. Integration with Google Translator ensures automatic detection and translation of input languages, enabling global applicability and enhanced reliability. For text analysis, the system utilizes BERT (Bidirectional Encoder Representations from Transformers), a large, pretrained model known for its strong contextual and semantic comprehension of human language. As the digital landscape rapidly evolves, precise identification of offensive content is becoming increasingly essential. Through this project, we are building robust, fair and impactful technology to foster safer online environments for all users, addressing this significant challenge head on.