Suraj Gupta
Management
July 2025
Social media is used unlike ever before in India and this has created an international level of interaction online but also abhorrent misuses of social media, such as hate speech, cyberbullying, and harassment towards individuals. This paper discusses how Artificial Intelligence (AI) can detect such abusive content automatically in an Indian multilingual and culturally diverse setting. Based on Natural Language Processing (NLP) and machine learning, such as Logistic Regression, LSTM, and BERT models, the given research assesses how effective AI is in handling abuse-filled material in English, Hindi, and Hinglish. Since BERT is a transformer type of model, it showed high accuracy (93.5) and context sensitivity and hence most reliable in detecting abuse. In spite of the obstacles arising due to the code-mixed languages and cultural nuances, the research signifies the possibility of AI-powered systems to achieve ethical, scalable, and real-time contents moderation and provide a safer Internet in India
139- 151
© Airo Journals 2021
