Starting in early 2023, students began spending less time on word problems and continued to spend about the same amount of time on graph problems. The gap widened every quarter. By the end of the study period, near the end of 2025, the average time spent on word problems had decreased by 31 percent among high school students and 27 percent among college students, from about four minutes per word problem to less than three. (Middle school students showed only a modest 9 percent decline, and fifth graders showed virtually none.)
Researchers believe those averages are declining because some students spend just a few seconds solving problems posed because they use AI to answer them.
The same pattern appeared on college-level tests. When exams were given unproctored, students spent much less time on problems raised after ChatGPT was released. During proctored exams, the time spent on posed problems returned to historical norms.
But time is only half the story. The most worrying finding is what happened to learning.
Many universities allow incoming students to retake placement exams after practicing more math in ALEKS, giving them the opportunity to qualify for a higher-level course. Before ChatGPT, that practice usually paid off. After ChatGPT, students correctly answered more problems posed during unproctored practice sessions, but performed substantially worse on those same types of problems when they later took a proctored placement test.
Historically, students correctly answered about 80 percent of these problems posed on proctored placement tests. After the introduction of ChatGPT, that figure fell to around 60 percent, a roughly 25 percent reduction in the odds of correctly responding to a posed issue.
On the contrary, performance on graphical problems did not decrease.
After the launch of ChatGPT, students performed worse on posed problems (susceptible to AI) during proctored exams, but answered more problems correctly in unproctored environments.
If students’ math skills had deteriorated overall due to learning loss from the pandemic, weaker preparation in high school, or digital distraction, graphing performance should have deteriorated as well. It wasn’t like that.
The study cannot definitively prove that the students were using AI. The researchers couldn’t see what else was happening on students’ screens outside of ALEKS. But it’s hard to think of any other explanation. The changes appeared only on problems that are easy to outsource to AI, disappeared under supervision, and grew steadily for almost three years.
“What makes me nervous is that these are not just issues raised,” Rismanchian told me. “This cognitive surrender could be happening in writing, science, everything.”
The newspaper, “Faster completion, less learning“, was published in June 2026 as a working paper and has not yet been peer-reviewed. Like any study, it does not resolve the questions of how many students are using AI in their schoolwork, whether it is harming learning and to what extent. But it adds to a growing body of evidence that generative AI is causing students to skip the brain work that leads to learning, and that this “cognitive surrender” is becoming commonplace.
A randomized experiment in Türkiye found that high school students who used AI to help them study mathematics ultimately lessons learned than the students who practiced without it. Anthropic, the creator of Claude, has separately reported that many college students appear to use AI to get answers and download cognitive work. Rismanchian’s previous research, published in March 2026, documented Worrying patterns of AI use in short-answer essays among undergraduates at a large California research university.
That doesn’t mean AI always undermines learning. Carefully designed AI tutors have improved student performance in controlled experiments by asking questions, personalizing instruction, and withholding answers until students reason to solve a problem. But using AI in this way should increase the time students spend on a problem, Rismanchian said. ALEKS data shows the opposite.
Rismanchian doesn’t think the answer is to simply ban AI. Instead, again, students need to value learning enough to resist the temptation to outsource it.
A recent RAND survey suggests that many already recognize the threat to their brains. Students report concerns that AI will be weakening your critical thinking skills while many of them admit to using it for schoolwork.
The students are not entirely to blame. Although many professors have warned students against using AI to complete classwork, universities themselves have embraced the technology and often give students free access to premium chatbots.
“I think we need to communicate to students that they need to value their learning,” Rismanchian said. “If ChatGPT does it for you, then you haven’t learned it.”
Rismanchian understands the temptation.
Rismanchian, an international student, started using ChatGPT to help polish his English in his papers. The ideas were still his. But after several months, he said, he noticed something disturbing.
“I realized I can’t write anymore,” he said. “I was losing my writing skills.”
So he stopped using AI to write.
He still uses it for coding.
