Don’t. Look. Away.
This is the advice you’re given if you use ProctorU, Proctorio or any of the other proctoring systems that have mushroomed in the wake of COVID-19 to manage students taking online tests. If exams are conducted online, the argument goes, we need a way to proctor them. And many companies are stepping forward to provide remote proctoring services to Utah State University, Brigham Young University, the University of Utah and schools across the nation.
The idea is simple. A student sits down to take an exam on a computer. They start up the proctoring app, which takes control of their web browser, their webcam and their desktop. An artificial intelligence-based, or AI, system records what the camera sees as well as the student interactions with the browser window, and flags any “suspicious” events. These events are then reviewed by an employee of the company and a summary of the analyses, both artificial and human, are reported to the instructor.
So, what’s considered suspicious? Well, if you looked away from the screen at any point while reading this op-ed, you just committed suspicious behavior. If your network connection paused because you lack a reliable internet connection, that’s considered suspicious, too. And don’t even think of going to the bathroom. The software flags any movement — even your roommate walking by — as suspect. Look down at your desk to check some allowable notes — also suspicious. All of this activity is detected by an AI system programmed to analyze video feeds and user clicks. It’s the same AI system designed for facial recognition to aid police that is criticized for its demonstrated bias against people with nonwhite skin colors.
Proctoring exams is a real challenge during COVID-19. Like all professors, I have to balance the trade-off between how much time I spend chasing down students who resort to cheating versus working to create an effective learning experience for my students. But what these systems do is much worse. They turn students into pre-cheaters, and their instructors into Big Brother.
Just like in the criminal legal system, a swarm of technology (including facial recognition) is being pitched to education institutions to detect impermissible activity. These systems are sold to school administrators with the promise of predicting illegal intent. They take cover under the guise of “we’re just flagging behavior, not making accusations.” And they prey on fears of increased “criminal behavior.” After all, if we can’t proctor exams, how will we prevent students from cheating?
But we must remember that our students are not cheaters, or even pre-cheaters. They are people struggling, like all of us, to manage their lives and education. Those of us who engage with students know how difficult it is for them to juggle childcare, jobs, health issues and their uncertain futures. Students must also retain privacy rights. They should be allowed to turn off their webcams to protect their privacy, which might be most of the time for some students. Schools could also promote remote meeting apps that use digital backgrounds to block everything beyond the student themself.
The proctoring apps being used at Utah universities provide none of this flexibility or privacy protection. If you don’t own a webcam, or can’t afford one, then you might fail the exam. If your test is interrupted by a crisis at your house, like a dog barking or a child needing attention, your “suspicious” behavior puts your grade at risk. Furthermore, the very design of these systems places the instructor in the role of cheating police. The atmosphere in the (virtual) classroom — already awkward — becomes even more tense. Just as we are struggling to build trust and connection with our students over a flickering screen and echoing audio connection, we introduce a system of surveillance that assumes our students are ill-prepared at best, or criminals at worst.
Teachers don’t want to be Big Brother. We want better tools to help our students thrive in this unprecedented environment. There is nothing wrong with using an honor code for test taking. If a few cheaters slip by, that’s a necessary price to pay for treating people like humans rather than criminals. And more compassion and support, rather than more invasive algorithmic surveillance, is what all of us need during these challenging times.
Suresh Venkatasubramanian is a professor at the School of Computing at the University of Utah and serves on the Utah Commission on Protecting Privacy and Preventing Discrimination. The viewpoints expressed by the author do not necessarily reflect the opinions of the University of Utah or the state of Utah.