In a world increasingly dependent on AI tools, a new study has raised red flags about the cognitive toll of overusing chatbots like ChatGPT. Researchers from MIT’s Media Lab, Wellesley College, and Massachusetts College of Art and Design conducted a four-month study revealing that frequent reliance on large language models (LLMs) can significantly diminish memory, reduce brain activity, and weaken a person’s sense of authorship and creativity.
The study involved 54 participants divided into three groups: one using ChatGPT, another using Google Search, and a control group writing without any digital assistance. Those using AI showed a notable drop in alpha band connectivity—a neurological marker for memory and language processing—compared to those who relied solely on their own thinking.
When asked to recall or quote from their previous essays, 83% of the ChatGPT users failed to remember key details or provide accurate citations. By the third session, most of their responses were nearly copy-pasted from AI outputs, with minimal edits.
Another fascinating observation was the shift in intellectual ownership. Unlike the control group—who strongly felt they authored their own work—AI users reported mixed and fragmented feelings. Some claimed full authorship, others partial, and many felt no ownership at all.
These findings open up a larger conversation about how generative AI tools may impact long-term human cognitive skills, and whether convenience is starting to replace real learning and creativity.