“The fact that students are misusing GenAI is a problem for assessment validity, and that’s a problem for the credibility of university credentials,” said Rene Kizilcec, associate professor of information science at Cornell University, one of three researchers who analyzed survey data from 95,000 students at 20 U.S. universities.
“About one-third regularly used generative AI such as ChatGPT or other models to produce text, video or code, when completing assignments, and 9 percent had used it to cheat,” according to the team, whose research comes amid wider concerns about AI “normalizing cheating at scale.”
“Even this early stage evidence shows that we have a very serious challenge on our hands, and universities need to address that,” warned Igor Chirikov of the University of California, Berkeley.
Published by the journal Science, the findings suggested that AI use and the likelihood of cheating vary depending on the field of study, with “significant” differences reported between disciplines.
While around 60 percent of computing students said they made at least monthly use of AI, compared to around a quarter of arts students, those in the science, technology, engineering and mathematics (STEM) arenas were less likely to cheat — or admit to doing so — by using AI.
“As we expect GenAI use among students to only grow, for better and worse, we also expect that GenAI misuse will grow, which is concerning,” Chirikov added, explaining that the survey analysis was done “to provide a more evidence-based approach” to understanding how students use and misuse AI.
©2026 dpa GmbH. Distributed by Tribune Content Agency, LLC.
