Abstract Book of the 9th World Conference on Research in Education
Year: 2025
[PDF]
Exploring Machine Learning Strategies to Mitigate Prerecorded Video Feeds Based Cheating in AI Proctoring Systems
Dr. Amadasun Osamuyimen
ABSTRACT:
Following the advancement in information Technology and the Covid 19 pandemic, has led to the wide spread adoption of AI Proctored examination systems in Educational space, To this end, protecting the integrity of online assessments has been facing new challenges. The development of reliable methods for detecting electronic cheating, notably the use of prerecorded video during examinations, has become essential with the rise of emerging cheating techniques. In this research, a thorough methodology for detecting prerecorded video usage in an AI-proctored test system is presented, In order to uncover suspicious activities connected with the use of prerecorded video feeds to outsmart the AI Proctoring systems. Advances in technology have led to more robust effective and efficient approaches incorporating deep learning models for real-time cheating detection from recorded video frames and speech (Kaddoura & Gumaei, 2022). Etc.). In this paper, we demonstrate how we used Lived and recorded video frames to extract relevant Datasets. Therefore, we presented an extensive experiments using cutting-edge machine learning techniques with curated datasets from Live video tagged as Non Cheating and Recorded video tagged as Cheating. A model was developed for detecting when recorded video is deployed for cheating during AI Proctoring exams. The findings indicate accurate electronic cheating detection that would improve academic evaluation integrity of AI proctoring systems.
Keywords: academic evaluation, prerecorded video feeds, Integrity of e-Assessment, AI proctored examination, electronic cheating