AI cheating in schools: how common is it really?
Most students believe AI-assisted cheating is now normal at their school. That's a different, and better-supported, claim than saying most assignments are AI-written.
What students and teens actually report
Pew Research's February 2026 survey of U.S. teens found 59% believe AI-assisted cheating has become a regular feature of student life. Common Sense Media's national survey of teenagers found a closely matching figure: six in ten students believe AI cheating happens regularly at their school, and a third say it happens extremely often.
Neither of those numbers measures actual cheating directly. They measure perception, which is a meaningful but different thing. A student can believe cheating is common without having done it themselves, and belief that "everyone's doing it" tends to run ahead of the actual rate in most cheating research, AI-related or otherwise.
What's driving the perception
The underlying usage rate gives some sense of why the perception is so widespread. RAND's American Youth Panel found 62% of students ages 12 to 29 had used AI for homework by December 2025, up from 48% seven months earlier. A 2026 Fortune report citing College Board data put the figure even higher among high schoolers specifically, at 84%, while finding only about three in ten schools have any formal rules governing that use.
That gap between how many students use AI and how few schools have rules for it is likely doing a lot of the work here. When there's no clear line between "getting help" and "cheating," students reasonably assume a lot of their classmates are landing on the permissive side of an undefined boundary. Common Sense Media's data adds a wrinkle to this: about half of students who used generative AI for schoolwork did so with a teacher's actual permission, meaning a meaningful share of what gets lumped into "AI cheating" in casual conversation is sanctioned use that simply isn't labeled that way.
Why detection numbers don't settle the question
Schools that try to measure cheating directly, rather than relying on perception surveys, mostly do it through AI-detection software, and those tools have real accuracy problems that distort any count based on them. Turnitin advertises 98% accuracy and a sub-1% false-positive rate, but that number reflects the company's own curated testing. Independent research, including a University of Chicago Booth working paper that stress-tested detectors on real student writing, found sentence-level false-positive rates closer to 4%. The same research found seven different AI detectors falsely flagged 61.3% of TOEFL essays, written by non-native English speakers, as AI-generated, compared to close to zero for native-English writing.
That means detector-based cheating counts run in both directions at once: they miss a meaningful share of actual AI-generated work that's been lightly edited to evade detection, while falsely flagging genuine student writing, especially from English-language learners, as AI-produced. Vanderbilt University disabled Turnitin's AI-detection feature entirely in 2023 over this exact bias risk. A school reporting "X% of assignments were flagged as AI-written" is reporting a number shaped as much by the detector's blind spots as by actual cheating.
What schools are actually doing about it
Given both the perception problem and the detection problem, most district policy has moved away from catch-and-punish toward assignment redesign and disclosure requirements, closer to how higher education has handled this. That includes in-class, unaided writing components for high-stakes assessments, oral defenses of written work, and requiring students to show drafts or process rather than only a final product. None of that eliminates AI-assisted cheating, but it reduces how much weight rests on a single, unreliable detector score.
The policy question of whether restricting or banning AI access reduces cheating is covered separately in should AI be banned in schools, where the real-world track record of NYC, LA, and Seattle's short-lived bans is relevant here too: none produced evidence that a network-level block meaningfully reduced AI use, since most homework happens off-network in the first place.
Small changes beat big bans
For parents and teachers weighing this, the fix is rarely a blanket ban, it's a few deliberate habits around how AI gets used. The free 5-Day AI Reset walks it in five short emails: measure where AI shows up, take one task back, name the urge, sort helpful use from skill-replacing use, and set rules that survive a bad week. No quitting AI, no guilt.
Frequently asked questions
What percentage of students cheat with AI?
There's no reliable direct measurement of actual cheating rates, since detection tools have significant accuracy problems. What's well documented is perception: roughly six in ten teens and students believe AI cheating is now a regular occurrence at their school.
Is AI cheating actually as common as students think?
Probably not exactly as common as the perception suggests. Usage rates for AI on homework are high (62% to 84% depending on the survey and age group), but a meaningful share of that use is sanctioned by teachers rather than hidden, and perception of peer cheating tends to run ahead of confirmed rates in most research on academic dishonesty.
Can AI detectors like Turnitin accurately catch cheating?
Not reliably. Vendor-claimed accuracy comes from internal testing, while independent research has found meaningfully higher false-positive rates and a specific bias against non-native English speakers, which has led some universities to disable AI detection entirely.
Do most schools have rules about AI use on assignments?
No. Recent reporting citing College Board data found only about three in ten schools have formal AI-use policies, even as the large majority of students report using AI for schoolwork.
What are schools doing instead of relying on detectors?
Many are redesigning assessments to include in-class unaided writing, oral defenses, or draft-and-process requirements, reducing how much a single detector score determines the outcome for any one student.
This article is informational and not a substitute for school-specific policy guidance.
Part of our guide to AI in schools: the real pros and cons. See also: why AI should not be used in schools · should AI be banned in schools.