The cheating generation is asking us to help them stop

Open this photo in gallery:

Bradley Richards teaches philosophy courses at Toronto Metropolitan University, and has co-created MATCHA Writer, software that prevents inappropriate AI use.

Ever since the launch of ChatGPT in 2022, and the rapid rise of large language models, students have been increasingly approaching me in class to report cheaters. It’s not fair, they say, for some of their peers to use prohibited LLMs while they toil. They worry that they’re getting pushed down the curve. But as well as fretting about their grades, which are crucial for admission to graduate programs and professional degrees, they also face a moral quandary.

These honest students want to learn. They want to grow personally and intellectually; they recognize that hard work and struggle are necessary for those goals. But as their achievements get lost in the sea of AI prose and undeserved A-pluses, they wonder if they must cheat too.

Cheating is not new. A pre-AI meta-analysis by Ball State University’s Bernard Whitley suggests that students usually cheat out of desperation. If students are balancing a demanding course load, a job, and social commitments, when a difficult assignment or test is due, finding shortcuts is tempting. But other studies – including a seminal one from Rutgers University’s Donald McCabe – suggest students are more apt to cheat when they think it is normal. With GenAI making cheating easier than ever, that’s exactly what’s happened – and increasingly LLMs are perceived as necessary for survival.

Opinion: Universities should stop treating AI as a cheating problem

Are enough students cheating that it’s really a problem? The answer is yes, but the data are vexed. A 2026 University of California article using data from the 2023-2024 academic year claims that a relatively small number of students are using AI to cheat: 9 per cent. This is much lower than many other self-report studies, mainly because it is already outdated given the meteoric rise of GenAI use. Despite using survey instruments designed to catch dishonest reports, these kinds of studies inherently face the problem that cheaters conceal their cheating. Some studies have shown that cheating rates are at least three times as high as self-reports.

The anecdotal consensus among professors is that we are inundated with assignments and tests completed with, or entirely by, LLMs. Reddit discussion boards are flooded with exasperated faculty members overrun by cheating. Likewise, widely reported incidents like one from Brown University suggest the vast majority of students cheat with LLMs when given the chance. After the average score on a take-home midterm in Roberto Serrano’s economics course at Brown was 96 per cent, he changed the final to an in-person exam. Eighteen students dropped the course and the average of the remaining 59 students was 48.6 per cent, with only a few earning similar scores on both tests.

To combat AI use, Princeton University has ended a 133-year exam practice. Since 1893, exams were unproctored, and students followed an honour code that included a pledge not to cheat. If cheating did occur, fellow students were expected to report it. Other schools are reviewing their honour codes; Middlebury College in Vermont became one of them after it received the highest self-report of violations – 65 per cent – since it began asking six years ago.

The exam environment is changing in Canada too; the Council of Ontario Universities reports interest in returning to oral exams, and suggests students be required to declare exactly how they used AI in each assignment, not just adhere to a generic code of conduct. (According to Rutgers’ Prof. McCabe, honour codes actually used to deter cheating, but their potency is waning.)

Some students that cheat with GenAI feel bad about it. KPMG reports that 73 per cent of Canadian postsecondary students admit to using AI in their school work, and 45 per cent say their first instinct is to use it when they get an assignment. And yet, more than half say they worry they are cheating and will get caught.

They’re also aware of the long-term consequences. Sixty-five per cent surveyed say their peers rely on AI to avoid critical thinking, and 48 per cent say their critical-thinking skills have deteriorated since they started using AI. According to recent research, they are right.

Despite the fact that they know it’s wrong and that AI is stunting their cognitive growth, my personal experience shows that they continue to cheat because they feel it’s normal, easy, and necessary for success in this new academic environment. Likely, that pressure won’t end when they enter the working world, where AI use is even more rampant.

Why the low-tech zine is making a comeback

According to some reports, since the launch of ChatGPT, the percentage of online content being generated by AI has climbed to 50 per cent. While some students and educators might say this is a reason to allow AI use for students, a 2026 University of British Columbia study on student perceptions of GenAI suggests this is not a consensus. Approximately half of the students surveyed disagree that UBC needs more GenAI in teaching and learning – and disagree also that they should be exposed to more GenAI for the sake of their careers. The study concludes that students are “not yet convinced of the long-term academic value of these tools.”

I am not anti-AI. I am pro-student. Cheating is becoming the new norm at universities, but it’s not because students want it to be that way. They know it’s wrong, and more importantly, they know it’s undermining their education. They’re asking us to help them stop.

Listen: Canadian professors on how AI is changing education