How Many Students Cheat With AI? Less Than They Think

How many students cheat with AI? Fewer than students themselves believe. In the first Dartmouth AI Survey, released October 30, 2026, 11% of US college students said they use AI to find answers to problem sets, while 49% thought their peers did. Eight percent said they paste AI text with little editing; 50% thought their peers did. That gap is not trivia. Decades of research on academic dishonesty find that what students believe their classmates do is one of the things that shapes what they do. “Everyone’s using it” is mostly a myth, and it is a myth that can make itself come true.

Key Takeaways

• Dartmouth’s 2026 national sample of 1,003 college students found self-reported AI cheating around 8–11%, while about half believed their peers did it.
• 39% of those students said they felt they had to use AI because their peers do, compared with 14% of incoming first-years surveyed before arriving on campus.
• Among US teens, 59% think AI cheating is a regular occurrence at their school (Pew, 2026), a perception, not a measurement.
• Stanford’s Challenge Success surveys found overall high school cheating rates of 60–70% before ChatGPT and roughly the same after it.
• Telling your teen what the data actually show, and that most peers disapprove of outsourcing whole assignments, is a concrete step that research supports.

What the Dartmouth AI Survey actually found

The Dartmouth AI Survey is a new yearly study of how college students use and think about AI, led by economist Bruce Sacerdote. The first results, posted on Dartmouth’s survey page, come from two groups: a weighted national sample of 1,003 US college students recruited through the College Pulse panel and surveyed September 10–23, 2026, and 1,533 Dartmouth students, including 332 incoming first-years.

The headline finding is the perception gap. Students were asked about their peers first, then about themselves.

Behavior (US college students)	Say they do it	Think their peers do it	Gap
Use AI to find problem-set answers	11%	49%	38 points
Copy and paste AI text with little editing	8%	50%	42 points
Dartmouth undergrads: use AI to draft essays	about 17%	63%	about 46 points

The survey also asked how students feel about all this. Seventy percent said take-home work has become unfair because of AI. Eighty-four percent said using AI for writing can hurt the development of writing skills. And 61% said they use AI to explain difficult concepts, which is the kind of use most teachers would welcome.

The detail that should get a parent’s attention is this one: 39% of the national sample said they feel they have to use AI because their peers do. Among Dartmouth’s incoming students surveyed before they arrived, the figure was 14%. The survey can’t tell us why it nearly triples, but the obvious candidate is the campus belief that everyone else is already doing it.

How many students cheat with AI, really? Testing the gap

This is the right question, and it deserves a straight answer: we can’t fully rule out that students underreport their own cheating. Self-reports are self-reports, and the Dartmouth team says so on its methods page.

But several things make a 40-point gap hard to explain by lying alone.

First, other measurements land in a similar place. Turnitin, which sells AI-detection software, told Education Week in 2024 that of more than 200 million writing assignments its tool reviewed in a year, about 3 in 100 were mostly AI-generated and about 1 in 10 showed some AI use. These are a vendor’s own numbers from its own detector, which has known accuracy limits, so treat them as rough. Still, nobody’s data shows half of all work being pasted from a chatbot.

Second, cheating rates didn’t jump when ChatGPT arrived. Stanford’s Denise Pope and Victor Lee described Challenge Success surveys of US high school students in which, for years before ChatGPT, 60 to 70% reported at least one cheating behavior in the previous month. In 2023, with new questions about AI added, that share “stayed about the same or even decreased slightly.” Lee was careful: this “may change as students become increasingly familiar with the technology.”

Read those two together. Cheating has always been common in some form, often small things like copying a homework answer. AI didn’t create it and, at least early on, didn’t obviously expand it. What AI did change is how visible and frightening it feels.

Third, perception gaps like this are an old, well-documented pattern. People routinely overestimate how often peers do things that are talked about, worried about, or covered in the news. AI cheating is all three.

Why believing “everyone cheats” makes cheating more likely

Here is the part that turns a fun statistic into something useful.

In a 1997 study published in Research in Higher Education, Donald McCabe and Linda Klebe Treviño surveyed students at nine state universities about academic dishonesty. Among the factors linked to cheating were the level of cheating among peers and peer disapproval of cheating. Peer disapproval was the strongest influence they found.

That gives you two levers, and the AI conversation tends to pull the wrong one. When teens hear constantly that everyone cheats with AI, two things shift at once: the perceived level of peer cheating goes up, and the perceived disapproval goes down. “If everyone does it, nobody really minds.”

The Dartmouth students describe exactly that pressure. And you can hear it in the Harvard Center for Digital Thriving’s interviews with teens, where one student said: “It sucks because I never was going to cheat, but now it feels like I have to.”

The Pew Research Center’s 2026 teen survey shows how widespread the belief already is among younger kids. Of US teens aged 13 to 17, 59% think using AI to cheat happens at least somewhat often at their school, and about a third say it happens extremely or very often. Among teens who have used chatbots for schoolwork, 76% say students at their school cheat with them at least sometimes.

Two honest limits. Pew measured belief, not behavior, so it doesn’t tell us how much cheating teens actually do. And the Dartmouth data are from college students; we don’t have an equivalent own-versus-peers comparison for middle schoolers yet.

Where the line actually sits for most students

Something else gets lost in the panic. Most students don’t think having AI do the whole job is acceptable.

Pope and Lee reported that in Challenge Success’s survey, many high schoolers thought chatbots should be acceptable for “starter” purposes, like explaining a new concept or generating ideas, “but the vast majority said that using a chatbot to write an entire paper should never be allowed.” Dartmouth’s 84% agreeing that AI writing can hurt writing skills points the same way.

Pope’s broader point is worth passing on. Students cheat for reasons that have “very little to do with their access to technology”: struggling with material and lacking help, too much homework and too little time, assignments that feel like busywork, and pressure not to disappoint their families. Students “are less likely to cheat when they feel a sense of belonging and connection at school, and when they find purpose and meaning in their classes.”

That moves the question from “how do I stop my kid using AI?” to “what pressure might push my kid toward it?” The second question is one a parent can do something about.

What to do with this at home

Correct the myth with numbers, not a lecture

Teens respond better to data than to moral warnings. Try: “There’s a survey of college students where about 1 in 10 said they used AI for problem-set answers, but half thought everyone else did. Why do you think people guess so high?” You’re not asking whether they cheat. You’re asking them to think about how beliefs spread, which is a better conversation and a more honest one.

Make disapproval visible, including theirs

Since peer disapproval was the strongest factor in McCabe and Treviño’s study, help your teen notice that most classmates do draw a line. Ask where they would draw it: Is AI explaining a math concept fine? Brainstorming? Rewriting a paragraph? Writing the whole essay? Most kids put the line somewhere sensible, and saying it out loud makes it theirs. Our guide on what parents should know before assuming a kid cheated has more on this conversation.

Look for the pressure underneath

If your child does lean on AI to get through assignments, Pope’s list is a diagnostic. Are they stuck on the material? Overloaded? Bored by work that feels pointless? Afraid of a bad grade at home? Each has a different fix: help or tutoring, a talk with a teacher about workload, or a change in how your family reacts to grades.

Ask the school to share real numbers

Dartmouth’s recommendation to colleges was to “lift the veil”: share what students actually report, alongside clear guidance. A school can do the same with an anonymous survey of its own students. If your PTA or school council is discussing AI, this is a practical, low-cost suggestion.

What not to do: tell your teen “everyone cheats now, so be careful”

It sounds protective. It reinforces the exact belief that raises pressure. Even “I know all your friends are using ChatGPT” quietly tells your kid that cheating is normal. Say what’s true instead: some kids do, most don’t think doing the whole assignment with AI is okay, and the ones who learn the material come out ahead. For a related trap, see why collaboration and cheating blur on social media.

What to Watch For Over the Next 3 Months

• Week 4: Listen for “everyone does it” or “it’s not fair, they all use AI.” That is the perception gap talking. Answer it with the numbers.
• Month 2 red flags: Your teen says they feel they have to use AI to keep up, grades that suddenly jump with no visible effort, or a child who can’t explain work they turned in. Any of these suggests pressure or overreliance worth a calm talk.
• Month 3 self-check: Does your teen know their school’s AI rules for each class? Can they tell you where they personally draw the line? Have you talked at least once about what pressures push kids to cut corners?

Frequently Asked Questions

How many students actually cheat with AI?

It depends on the measure. In the 2026 Dartmouth AI Survey, 11% of US college students said they use AI to get problem-set answers and 8% paste AI text with little editing. Turnitin reported in 2024 that about 3% of assignments it scanned were mostly AI-written. High school cheating overall has run 60–70% for years, mostly in smaller forms.

Why do students think everyone else cheats?

People tend to overestimate behaviors that get a lot of attention and worry. AI cheating is discussed constantly by teachers, parents and the media, so it feels universal. The Dartmouth survey found roughly 40-point gaps between what students reported doing and what they believed peers did.

Does believing others cheat make my teen more likely to cheat?

Research suggests it can. A 1997 multicampus study by McCabe and Treviño found cheating was linked to the level of cheating among peers and to peer disapproval, with disapproval the strongest factor. In the Dartmouth survey, 39% of college students said they felt they had to use AI because their peers did.

Has AI made cheating more common in high school?

Early evidence says not much. Stanford’s Challenge Success surveys found 60–70% of high schoolers reported some cheating behavior before ChatGPT, and the share stayed about the same or dipped slightly in 2023. The researchers cautioned that this could change as students get more familiar with AI.

Is using AI for homework always cheating?

No. It depends on the assignment and the teacher’s rules. Using AI to explain a concept or check understanding is often allowed; having it write the whole assignment usually isn’t. Most students agree: Stanford’s researchers found the vast majority said a chatbot writing an entire paper should never be allowed.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. Sacerdote, B., & Parris-Bacon, M. (2026). “Dartmouth AI Survey 2026: How college students think, learn, and work with AI.” Dartmouth College. https://ai.dartmouth.edu/dartmouth-ai-survey-2026
2. Dartmouth College. (2026, October 30). “Survey reveals what students think about AI.” Phys.org. https://phys.org/news/2026-10-survey-reveals-students-ai.html
3. Pew Research Center. (2026, February 24). “How Teens Use and View AI.” https://www.pewresearch.org/internet/2026/02/24/how-teens-use-and-view-ai/
4. Stanford Graduate School of Education. (2023). “What do AI chatbots really mean for students and cheating?” Interview with Denise Pope and Victor Lee. https://ed.stanford.edu/news/what-do-ai-chatbots-really-mean-students-and-cheating
5. McCabe, D. L., & Treviño, L. K. (1997). “Individual and Contextual Influences on Academic Dishonesty: A Multicampus Investigation.” Research in Higher Education, 38(3), 379–396. https://link.springer.com/article/10.1023/A:1024954224675
6. Prothero, A. (2024, April 25). “New Data Reveal How Many Students Are Using AI to Cheat.” Education Week. https://www.edweek.org/technology/new-data-reveal-how-many-students-are-using-ai-to-cheat/2024/04