Increasingly, teachers and schools fretting over students using artificial intelligence to complete their assignments are turning to AI detectors to catch would-be cheaters.
That may sound like a smart countermeasure to a pernicious tech shortcut. But it’s putting academic success in peril for students like me.
As a high school junior in the rigorous International Baccalaureate program, I am fortunate to be surrounded by students who take their educations seriously.
We get our work done, and we get it done right. If students are caught cheating, they risk not receiving their IB diploma — even after two years of grueling college-level work.
Now, you may ask yourself: What should students worry about AI detectors if they’re following the basic rules of academic integrity?
Unfortunately, it’s not that simple. AI detection tools are grossly inaccurate.
In explaining its policy discouraging their use, the University of San Diego School of Law noted that “multiple studies have shown that AI detectors were ‘neither accurate nor reliable,’ producing a high number of both false positives and false negatives.”
Turnitin, a popular detector, says it has a false positive rate below 1 percent, but emphasizes that “the final decision on whether any misconduct has occurred rests with the reviewer/instructor.”
In 2023 when it launched, a Washington Post product test found a lack of reliability. A 2025 study from researchers at Sam Houston State University that reviewed the efficacy of Turnitin, GPTZero and other AI detectors concluded that “effectiveness significantly decreases in real-world academic settings.”
This semester, I saw this firsthand. After my classmates and I had worked on an assignment for two weeks, my teacher informed us that Turnitin had found that 10 of us submitted work that was either close to or 100 percent AI-generated. This was in a class of 23 students.
My paper wasn’t flagged. But authors of the 10 that were received a stern rebuke: The teacher instructed them to resubmit the assignment with a lower detection score or get a zero.
My classmates offered evidence that the work was authentic and that AI detectors are inaccurate, but the instructor stood by Turnitin’s assessment.
My fellow students, meanwhile, were left to anxiously wonder how to avoid false accusations of cheating.
I’m sure students everywhere can sympathize with the conundrum educators are facing. I certainly do.
Academic cheating is a serious matter that AI has suddenly made diabolically easy to pull off. An infraction can severely damage a student’s academic prospects and, with it, career success. In Advanced Placement courses, “unacceptable use of AI” can result in an instant zero.
Given such consequences, even a 1 percent false-positive rate is too much — what’s remotely acceptable about 1 in 100 students having their future prospects harmed for something they didn’t do?
