A new study from MIT's Media Lab suggests students who use ChatGPT for essay writing show measurably lower brain activity than peers who work without artificial intelligence assistance, raising fresh concerns about the cognitive costs of AI reliance in education.
The research, released Monday ahead of peer review, tracked 54 participants over four months as they wrote SAT-style essays under three conditions: using ChatGPT, Google search, or no digital tools. EEG monitoring revealed ChatGPT users exhibited the weakest neural engagement across 32 brain regions, while those working without technological support demonstrated the strongest cognitive activity.
Researchers found ChatGPT users "consistently underperformed at neural, linguistic, and behavioral levels" compared to other groups1. By their third essay, many participants had shifted to copying and pasting AI-generated content with minimal editing1. Two English teachers who evaluated the essays described the ChatGPT-assisted work as largely "soulless"1.
Students working without digital assistance showed the highest neural connectivity in brain regions associated with creativity, memory processing, and semantic understanding1. When these participants were later introduced to ChatGPT in the fourth session, their brain activity actually increased rather than decreased, suggesting prior cognitive development enhanced their ability to collaborate with AI tools2.
Lead researcher Nataliya Kosmyna, a research scientist at MIT's Media Lab since 2021, chose to release findings before completing peer review due to concerns about educational policy decisions1. "What really motivated me to put it out now before waiting for a full peer review is that I am afraid in 6-8 months, there will be some policymaker who decides, 'let's do GPT kindergarten,'" Kosmyna told Time magazine1. "I think that would be absolutely bad and detrimental. Developing brains are at the highest risk."
The study introduces the concept of "cognitive debt" – a deficit that accumulates when over-reliance on AI tools reduces mental effort invested in learning2.
The findings emerge as school districts nationwide grapple with ChatGPT policies. New York City Public Schools initially banned the tool before reversing course within four months, while other districts maintain restrictions or conduct ongoing reviews1. A UNESCO survey found less than 10% of schools have implemented formal guidance on generative AI use1.
Meanwhile, a separate systematic review published this month in Computers and Education found ChatGPT can foster student engagement when properly integrated2. The conflicting research underscores the complexity facing educators as AI tools become ubiquitous.
"Reliance on AI systems can lead to a passive approach and diminished activation of critical thinking skills," the MIT study concluded3.
