How common is AI cheating in college, really? New research suggests it’s both less widespread than many headlines imply and concentrated among the heaviest users. A large survey of more than 95,000 undergraduates at 20 public research universities, conducted in spring 2024 and analyzed in research Cornell published in 2026, estimated that about 9% of students had used AI to cheat. Among students who use generative AI daily, that rate rose to about 26%, compared with 7% among monthly users.
This article summarizes recently published research and news reports as of late 2026. New studies appear regularly, and different surveys measure cheating in different ways, so the figures below are best read alongside each study’s methods.
Most Students Use AI, and a Third Use It Regularly
According to reporting on the survey, about two-thirds of the students had used AI for classwork, and about one-third used it regularly. Regular use varied by group: Cornell’s summary reported that 33% of female students used generative AI regularly, compared with 45% of male students, and that students from underrepresented racial minority groups had lower rates of regular use, 29%, compared with 39% of white and Asian students.
Those numbers match the broader trend described in our earlier roundup of college cheating and AI statistics, and they help explain why AI use allegations are increasing even among students who didn’t intend to cheat.
How the Researchers Measured Cheating
Students often hesitate to admit breaking the rules, even in anonymous surveys. To get a more honest estimate, the researchers used a statistical technique called a list experiment, which estimates sensitive behavior without asking any individual to confirm it directly. That approach produced the roughly 9% estimate. The researchers noted that this was lower than anecdotal reports had suggested.
Cheating Rates Vary by Field
Reporting on the survey also found differences by major: while about 9% of students overall reported cheating on exams using AI, the figure rose to 13% for business and management students and 17% for economics or journalism students.
What’s Happening in Other Countries
Similar trends are showing up abroad. In the United Kingdom, the Higher Education Policy Institute’s 2026 Student Generative AI Survey found that 12% of students had included AI-generated text directly in a piece of assessed work, up from 8% in 2025, and that 15% believed such use was acceptable. At Oxford, the student newspaper Cherwell reported, based on Freedom of Information requests, that investigations into unauthorized AI use rose to 35 in the 2025–26 academic year, from 24 the year before.
What This Means for Accused Students
The research cuts in two directions. AI misuse is real and measurable, which is why universities are investing in detection and enforcement. But the data also suggests that most students, including most AI users, aren’t cheating, which matters because AI detection tools can misidentify honest work, as explained in AI detectors, cheating, and plagiarism: what every student needs to know. A rising number of investigations means more students facing accusations, and some of those students will be wrongly accused.
Frequently Asked Questions
What percentage of college students cheat with AI?
A survey of more than 95,000 undergraduates at 20 public research universities estimated about 9%, using a method designed to reduce underreporting.
Are daily AI users more likely to cheat?
Yes. The same research found cheating rates of about 26% among daily generative AI users, compared with about 7% among monthly users.
How many college students use AI for schoolwork?
In the same survey, about two-thirds of students had used AI for classwork, and about one-third used it regularly.
Are AI cheating investigations increasing?
Where data is available, yes. At Oxford, for example, investigations into unauthorized AI use rose from 24 in 2024–25 to 35 in 2025–26, according to the student newspaper Cherwell.
Getting Help With an AI Accusation
As investigations rise, more students are being asked to explain their work or answer an AI-related charge. Richard Asselta, a student defense attorney, represents students at Florida universities and advises students facing AI-related academic misconduct allegations nationwide. He helps students understand the evidence against them, prepare their response, and work through their school’s process. For definitions of common terms in these cases, see our academic misconduct glossary. Call 855-338-5299 or learn more about Richard Asselta.