Question Bank Plagiarism Checker Module

In the digital age of education, ensuring the originality and integrity of question banks is a pressing challenge, especially with the widespread reuse and paraphrasing of content. Manual plagiarism detection methods are inefficient at scale and often fail to identify semantically similar or reworded questions. This paper presents an AI-Based Question Bank Plagiarism Checker Module that leverages advanced Natural Language Processing (NLP) techniques to detect duplicated and paraphrased academic questions. The system utilizes transformer-based sentence embedding models such as Sentence-BERT and SimCSE to convert textual content into semantic vectors. Cosine similarity metrics are then applied to determine the degree of similarity between questions. The platform supports uploads in PDF, DOCX, and TXT formats and includes metadata classification (subject, grade, board, question type). A React-based frontend combined with a Node.js backend enables real-time detection, file-to-file comparison, and automated report generation. Administrators have access to dashboards for user activity tracking, file history management, and system insights. By automating semantic-level plagiarism detection, the system significantly improves academic content quality and supports fair assessment practices.