AI-Powered Online Exam Proctoring System

Abstract

The AI-Powered Online Exam Proctoring System is designed to ensure secure, fair, and reliable online examinations in remote environments. With the increasing use of digital learning platforms, maintaining academic integrity during online exams has become a major challenge. This system uses Artificial Intelligence, Computer Vision, Audio Analysis, and Natural Language Processing to monitor candidate behavior in real time and detect suspicious activities. It verifies the presence of the candidate through webcam-based face detection and identifies violations such as multiple faces or absence during the exam. The system also analyzes audio to detect background conversations using speech-to-text technology. Additionally, it captures screenshots, audio, and video recordings as evidence for review. An AIbased answer evaluation module is included to assess descriptive responses using keyword and semantic similarity techniques. All activities are analyzed to generate an anomaly score, helping administrators easily identify potential cheating cases. Overall, the system enhances the security, transparency, and efficiency of online examinations while reducing the need for manual invigilation.