Proctorio AI Exam Proctoring System Flagged Students of Color Disproportionately

During the COVID-19 pandemic, Proctorio became one of the most widely adopted AI-powered exam proctoring systems as universities shifted to remote learning. The software used facial recognition, eye-tracking, and behavioral analysis to monitor students during online exams, automatically flagging suspicious behavior for instructor review. However, research and student reports revealed systematic bias against students of color, particularly those with darker skin tones. The bias manifested in multiple ways. Proctorio's facial recognition system had difficulty detecting and tracking faces with darker skin tones, leading to higher rates of flagging for 'face not detected' violations. The gaze-tracking technology similarly struggled with darker eyes, incorrectly interpreting normal eye movements as suspicious looking away from the screen. Students in non-traditional testing environments, such as shared living spaces common among lower-income students, were also disproportionately flagged for background noise or movement. Documented cases emerged across multiple universities. At the University of Illinois, Black students reported being flagged at significantly higher rates than white students for identical behaviors. Miami University saw similar patterns, with students of color receiving academic integrity violations at disproportionate rates. The University of Vermont conducted an internal analysis that confirmed the bias patterns, leading them to discontinue Proctorio in 2021. Student advocacy groups and researchers documented these patterns through surveys and data requests. A study by the Algorithmic Justice League found that Proctorio's facial recognition component had error rates of up to 34% for Black women compared to less than 1% for white men. Students reported psychological distress from being repeatedly flagged, with some developing anxiety about taking exams and others facing academic probation or course failures based on AI flags. Proctorio's response to criticism was notably aggressive. The company filed DMCA takedown requests against researchers and critics, including attempting to remove academic papers and student testimonials from the internet. CEO Mike Olsen personally sued University of British Columbia student Ian Linkletter for $90,000 after Linkletter posted concerns about the software on Twitter. These legal tactics drew widespread condemnation from academic freedom advocates. By 2022, over 20 universities had discontinued their contracts with Proctorio, citing bias concerns, student complaints, and privacy issues. Some institutions implemented human review requirements for all AI flags, while others moved away from automated proctoring entirely. The incident highlighted the risks of deploying biased AI systems in high-stakes educational settings and the importance of algorithmic auditing in EdTech products.