Detecting Cheating in Proctored Tests Through Pupil Tracking

The growing prevalence of online assessments - ranging from academic exams to technical interviews - has been accompanied by a corresponding rise in dishonest behavior. Traditional methods of preventing cheating, such as live human proctoring or lockdown browsers, are either too invasive, expensive, or easy to circumvent. This report explores an innovative approach that leverages eye-tracking technology to detect cheating behaviors in a more scalable, less intrusive way.

Using Pupil Core, an open-source eye-tracking platform that combines real-time gaze detection with wearable hardware, this project investigates whether tracking gaze and pupil behavior can reliably distinguish between cheating and non-cheating scenarios. The goal is to evaluate if gaze data can serve as a behavioral biomarker for academic dishonesty in remote settings.