This article explores growing concerns among educators and researchers that students who rely on ChatGPT for writing assignments may engage in less critical thinking and learning. Citing a recent MIT study and classroom experiences, it highlights how AI-generated work often lacks originality and insight, raising questions about the impact of AI tools on education.
The growing use of ChatGPT in classrooms is raising alarm bells among educators and researchers who fear that students relying on the AI tool may be learning less—and thinking less critically.
Jocelyn Leitzinger, a professor at the University of Illinois Chicago, has seen this firsthand. In her undergraduate business and society course, students were tasked with writing about personal experiences with discrimination. But many of the essays repeated the same fictional character—“Sally”—as the victim.
"It was very clear that ChatGPT had decided Sally was a common woman's name," Leitzinger said. "They weren’t writing from their own lives. These were AI-generated stories."
Leitzinger estimated that nearly half of her 180 students misused ChatGPT during the semester—including for essays on the ethics of artificial intelligence, which she described as both “ironic” and “mind-boggling.”
Her observations align with the findings of a recent study from MIT researchers, which suggests that students who use ChatGPT to write essays may engage in less critical thinking.
Table of contents [Show]
In the small, preprint study—yet to be peer-reviewed—54 adult students from the Boston area were split into three groups: one used ChatGPT to write 20-minute essays, another used a search engine, and the third relied only on their own knowledge. Brain activity was monitored using EEG headsets, and teachers graded the essays.
The results? Students using ChatGPT scored lower than their peers across all evaluation criteria. Their EEG readings also showed reduced brain connectivity, indicating less cognitive engagement. Alarmingly, over 80% of the ChatGPT users couldn’t recall any details from their own essays shortly after writing them. In contrast, only about 10% of the other groups had the same problem.
By the third round of the study, many ChatGPT users were simply copying and pasting responses, researchers observed. Teachers described these AI-generated essays as grammatically sound but devoid of insight, creativity, or personal voice—what one called “soulless.”
Still, the study’s lead author, Nataliya Kosmyna, cautioned against overgeneralizing the findings. She emphasized that the study was small and experimental. In a later round where the brain-only group was given ChatGPT as a writing tool, their brain activity showed higher levels of neural engagement—suggesting that thoughtful, guided use of AI might enhance learning.
“This is not about AI making us stupid,” Kosmyna said. “It’s about understanding how we use these tools and what effect that has.”
Ashley Juavinett, a neuroscientist at UC San Diego who was not involved in the research, criticized sensationalist headlines and noted the study’s limitations. “The methodology does not support broad claims about the neurological impact of large language models like ChatGPT,” she said.
Back in the classroom, Leitzinger says she’s watched student writing decline in both grammatical error and original thought since ChatGPT’s release in 2022. Some students don’t even bother to reformat the AI’s output before submitting it.
Despite these concerns, she also expressed empathy. Students are often caught in a confusing gray zone—some instructors encourage AI use, while others ban it entirely.
“It’s like when calculators were introduced—we had to rethink how we teach,” she said. But she warned that AI can let students skip foundational steps in learning, like developing a basic understanding before writing.
One British university student, who asked to remain anonymous, acknowledged the value of ChatGPT for generating ideas and organizing notes, but drew a line. “Using it to write your work for you? That defeats the point of being at university,” he said.
The concerns extend beyond the classroom. Academic journals are now grappling with a flood of AI-generated submissions. The publishing world is also feeling the impact, with some startups aiming to produce thousands of AI-written books annually.
Leitzinger summarized the core issue with a pointed question: “Writing is thinking, thinking is writing. If we remove that process—what does that mean for thinking itself?”
As AI tools become more integrated into everyday learning, educators and researchers alike are urging deeper reflection—not just on how they’re used, but on what could be lost in the process.
