ChatGPT use linked to weaker memory and brain connectivity: MIT
What's the story
A recent study by researchers at the Massachusetts Institute of Technology (MIT) has raised concerns over the cognitive impact of generative AI tools like ChatGPT. The 2025 research, titled Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task, examines how these tools affect human cognition, memory, and learning processes. The study looked at how reliance on AI influences cognitive engagement with writing tasks and whether frequent use could lead to "cognitive debt."
Research methodology
How the research was conducted
The MIT study involved about 54 participants aged between 18 and 39 years.
They were divided into three groups: one that used ChatGPT (LLM group), another that relied on Google Search (search engine group), and a third that depended solely on their own thinking (brain-only group).
Over several months, all participants completed multiple essay-writing tasks while their brain activity was continuously monitored using electroencephalography (EEG).
Results analysis
Brain activity levels varied significantly across groups
The study found that participants who wrote essays without any AI tools showed the highest and most widespread neural connectivity.
Those using search engines showed moderate engagement, while the ChatGPT group had lower levels of brain connectivity during the task.
There were also stark differences in memory and recall, with those using ChatGPT being less able to accurately remember or quote their own essays.
This suggests a reduced depth of cognitive processing during writing tasks when relying on AI assistance.
User behavior
Understanding 'cognitive debt'
The term "cognitive debt" in the study refers to the potential long-term cost of outsourcing thinking tasks to AI systems.
The researchers believe that continued reliance on large language models (LLMs) could lead to reduced independent engagement, weaker memory formation, and a lower sense of ownership over written work.
They also observed behavioral patterns among AI users such as increased reliance on copy-paste strategies and less variation in writing style across essays over time.
Research constraints
Study is preliminary, critics caution
Despite its significance, the MIT study is preliminary and based on a small sample size.
Critics have also argued that EEG-based measures of brain connectivity don't necessarily indicate long-term cognitive decline or reduced intelligence.
The authors and external researchers have warned against overinterpreting these findings as definitive evidence of harm.
Experts advise a balanced approach to AI usage: make your brain try the task before you ask the machine.