AI-Based Detection of Caste and Religion-Based Inequality with Legal Awareness System in India

This paper presents an AI-based system for detecting caste and religion-based discrimination in online text while promoting legal awareness. The system is implemented using Python and Streamlit and applies rule-based keyword scoring, text preprocessing, and severity classification to detect harmful content. The platform provides real-time visual alerts, analytical dashboards, and persistent logging through an SQLite database. It also integrates relevant legal information such as Articles 15 and 17 of the Indian Constitution and IPC Section 153A. Experimental evaluation on 250 text samples achieved an overall F1-score of 0.908, demonstrating reliable detection of discriminatory language and supporting ethical governance of digital platforms.