SMART PROCTOR CHATBOT FOR REAL TIME ONLINE EXAM SURVEILLANCE

The rapid shift towards online education has amplified the need for secure and scalable examination environments.Traditional human-invigilated proctoring methods are insufficient for maintaining academic integrity in remotesettings, often suffering from scalability, bias, and inefficiency. To address these challenges, this paper proposes theSmart Proctor Chatbot, an AI-driven real-time surveillance system designed to detect and prevent cheating duringonline exams. The system integrates advanced computer vision and machine learning technologies, includingYOLOv8 for object detection, MediaPipe for head pose estimation, and Roboflow for material recognition. Throughcontinuous webcam and microphone monitoring, the chatbot identifies suspicious behaviors such as multiple faces,mobile phone usage, and gaze deviation. Upon detecting anomalies, the system generates real-time alerts and logsincidents for administrative review. Experimental evaluation demonstrates the system's effectiveness in minimizingmalpractice while reducing reliance on human proctors. The Smart Proctor Chatbot offers a scalable, unbiased, andprivacy-conscious solution for secure online assessments, contributing to the future of trustworthy digital education.