ChatGPT is widely admired for its speed, intelligence, and ability to answer almost any question, making it a trusted companion for millions across the globe. However, a new and eye-opening study from MIT’s Media Lab suggests that this digital assistant might be doing more harm than good—particularly when it comes to our brain function. The research indicates that overreliance on AI tools like ChatGPT could be quietly impairing memory, critical thinking, and even fundamental brain activity.
The study, titled “The Cognitive Cost of Using LLMs,” was recently published on arXiv and delves deep into how large language models (LLMs), especially ChatGPT, impact our capacity to think, learn, and remember. The researchers sought to uncover whether depending on such tools might be coming at the price of our mental sharpness.
To examine the idea of a “cognitive cost,” the MIT scientists monitored 54 students over a four-month period. They used electroencephalography (EEG) headsets to measure their brain activity and divided them into three groups. One group regularly used ChatGPT to complete their assignments. Another group relied on Google for help. The third group was dubbed the “Brain-only” group and received no external assistance, using only their own reasoning and knowledge to complete tasks.
At first, the ChatGPT group seemed to have the advantage. Their results came in faster and with less effort, giving the impression that AI had improved their productivity. But over time, the findings revealed a concerning downside. Students who used ChatGPT for writing essays demonstrated worse memory retention, less brain engagement, and lower test scores when compared to those who didn’t use AI at all. The research team summed up their findings by stating, “The LLM group’s participants performed worse than their counterparts in the Brain-only group at all levels: neural, linguistic, and scoring.”
Interestingly, even the group that used Google outperformed the ChatGPT group. Although they did not show the same high levels of brain activity as the Brain-only group, the Google users still demonstrated moderate engagement and came up with more thoughtful responses than their ChatGPT-using peers. On the other hand, those relying solely on their brains produced original content and more insightful analysis.
The negative effects of ChatGPT extended beyond the initial use. Even when participants from the ChatGPT group were later asked to complete tasks without any help, their brain activity remained lower than the other groups. In contrast, participants from the Google and Brain-only groups showed increased brain activity as they adjusted to new methods. This led researchers to believe that regular use of ChatGPT may alter not just how we think, but whether we actively think at all.
One of the most sobering insights from the study is how AI usage fosters mental passivity. ChatGPT users reported that the tool made it easier to get information and complete assignments, but that same ease appeared to reduce their motivation to question or critically analyze what the AI produced. As the researchers explained, “This convenience came at a cognitive cost, diminishing users’ inclination to critically evaluate the LLM’s output or ‘opinions’.”
Another troubling point raised in the study concerns algorithmic bias. When ChatGPT provides information, it does so based on patterns in the data it was trained on—not necessarily what is factual or intellectually valuable. These patterns are often influenced by the goals of the companies behind the AI models, including profit motives and stakeholder priorities. The researchers suggest this creates a more advanced form of the echo chamber effect, where AI doesn’t just repeat what it’s learned—it shapes user thinking by presenting its output as truth. Rather than developing independent thought, users risk falling into a pattern of accepting AI-generated answers without scrutiny.
As AI tools continue to become a part of daily life—from drafting emails to writing essays and doing research—this study serves as a wakeup call for everyone, especially students, educators, and working professionals. While it’s tempting to see ChatGPT as a powerful partner in productivity, the research emphasizes that it shouldn’t become a substitute for genuine mental effort.
The researchers emphasized that caution is necessary as these tools become even more advanced. Users need to stay aware of the possible negative effects on mental processes. They warned, “In a world where convenience is king, critical thinking might just be the first casualty.”
The broader message of the study is clear: while AI is an extraordinary technological advancement, it’s not without consequences. Using tools like ChatGPT without awareness or limits may diminish key cognitive skills over time. Instead of empowering users, it can encourage mental shortcuts and dependency, weakening the brain’s natural ability to process, question, and remember.
Students who may turn to ChatGPT to complete an essay more quickly should be mindful of what they might be giving up in return. It’s not just about the quality of the content or getting a better grade—it’s about how their brain functions are being shaped by the tools they rely on. As the study shows, those who chose not to rely on AI ultimately emerged with stronger cognitive engagement and better overall performance.
In light of these findings, educators may need to rethink how AI tools are introduced in classrooms. While banning them outright may not be realistic, guiding students on how to use AI mindfully and in moderation could help preserve critical thinking. Professionals in creative or analytical fields should also reflect on their own usage. If convenience leads to complacency, the long-term cost could outweigh the short-term benefits.
In the end, the message from MIT’s study is not to avoid AI altogether, but to recognize its potential impact on how we think and learn. As AI continues to evolve, our understanding of its effects on human cognition must keep pace. Otherwise, we may find that in solving our problems more quickly, we’ve inadvertently slowed down our minds.
As the researchers put it, “This convenience came at a cognitive cost.” And as AI becomes more deeply woven into the fabric of daily life, that cost may become one we can no longer afford to ignore.