Artificial intelligence is often touted as a way to make teachers more efficient by helping them write lesson plans, generate teaching materials, and provide feedback to students in seconds. But one of the first randomized trials testing AI in real classrooms found that it could also harm learning. Students whose teachers had access to an AI teaching assistant felt less motivated to learn.
The damage was especially pronounced among students whose teachers were already weaker instructors, as measured by their performance before the experiment began. Their students also performed lower on standardized final exams, the researchers found.
“Teachers, just like students or coders, could use AI as a crutch,” said Alp Sungu, lead author of the study and assistant professor at the Wharton School at the University of Pennsylvania. “Instead of doing the actual work, they use AI to delegate the task, which lowers the quality of their teaching.”
A draft of the study, “Generative AI can harm teaching“, was published online in June and has not yet been published in a peer-reviewed journal. It echoes Sungu’s widely discussed research from 2024 on how students’ use of AI harms learning.
“Students are using AI as an answering machine, not as a learning tool, and so it is detrimental to learning,” Sungu said. “Here I think teachers are potentially using AI as a machine to generate material for homework, class notes, lesson plans, syllabuses. Instead of improving their own results, they are using AI as a replacement with very minimal interaction, and therefore the quality of the results is not good enough.”
Related: Faster Solutions, Lower Test Scores: How AI Is Eroding Math Skills
Sungu’s experiment, conducted with fellow researchers at the University of Pennsylvania including educational psychologist Angela Duckworth, followed 193 teachers and more than 2,800 middle and high school students at a chain of private schools in Turkey in the spring of 2025.
Teachers were randomly assigned to either have access to a ChatGPT-based teaching assistant tailored to the Turkish national curriculum or to continue teaching as usual. For 10 weeks, teachers primarily used the tool to generate class notes, assignments and exams.
Students whose teachers had access to the AI tool rated their lessons as less enjoyable, less interesting, and less important than students in the control group. The decline in intrinsic motivation was modest, but greater among students of teachers who were already heavy AI users before the experiment began.
The average academic result has not changed overall. But among teachers whose students had lower grades before the experiment — an indicator of lower-performing teachers — achievement and student confidence both declined. Academic achievement was measured using externally administered standardized exams, thus ruling out the possibility that these teachers had different grading standards.
The study cannot explain exactly why the quality of teaching has deteriorated. The researchers did not observe classrooms or analyze the AI-generated materials used by teachers. But Sungu suspects that teachers may have given up one of their most effective tools.
“When you start using AI-generated material, you lose your personal voice,” Sungu said. “It might be pretty good technically, but it doesn’t really reflect your own style. If everything is very uniform, it just becomes a little more boring.”
One possible explanation for the lower academic achievement among students of low-performing teachers, Sungu said, is that stronger teachers treat the AI results as a first draft, revising and adapting them for their classes. Weaker teachers, he suspects, might be more likely to use AI-generated materials as is.
Related: AI gives black students more praise and less criticism
This study does not constitute a clear comparison between teaching with and without AI. Teachers in the control group were free to use other AI tools, which is a comparison between access to a personalized AI assistant and anything teachers choose to do themselves. If anything, Sungu said, these results might underestimate the risks of teachers relying heavily on AI-generated materials.
Yet Sungu warns that would be a mistake to conclude that “AI is terrible and will ruin education”. He sees another lesson: Access to AI technology alone does not improve teaching.
The challenge is to help teachers use AI in ways that preserve human judgment and creativity. This will require teacher training programs, safeguards and better interfaces.
“Right now, there’s cause for concern about how organically teachers are using it,” he said.
Sungu says he personally uses AI in his university teaching to create interactive games and surveys that would otherwise take too long to create. “When I first get the result, it looks great,” he said. “And then, if I don’t delve into it, the examples, the numbers don’t make any sense. I end up spending so much time improving the result or calibrating it according to my class.”
“It’s not a waste of time,” he said.
Contact staffwriter Jill Barshay at 212-678-3595, jillbarshay.35 on Signal or barshay@hechingerreport.org.
This story on AI in education was produced by The Hechinger reportan independent, nonprofit news organization that covers education. Register for Proof points and others Hechinger Newsletters.
The article Teachers save time with AI. Their students could pay the price, which appeared first in the Hechinger report.
