The Massachusetts Institute of Technology (Mit) published one study To understand if and how the use ofartificial intelligence influence thebrain during writing. Many have read this job as “Chatgpt is making us stupid”, but that’s not the case! The results show that Write the brain without activating morebut they do not indicate any loss of intellectual skills when using it. “No stupidity, no brains on vacation,” said the project manager, Nataliya Kosmyna. What the study suggests to us is now Use chatgpt with moderation And to always try to write the texts independently, before turning to the AI.
The work, published on 10 June 2025, not yet officially overhauled From other experts, he attracted a lot of attention. And, although the authors themselves have repeatedly invited to avoid simplifications such as “Chatgpt is making us stupid”, the news has already circulated with alarmistic tones.
In this article we deepen the study, its results and what are the critical points.
The Study of the EMIT: the impact of chatgpt on our brain
The study, conducted by the Massachusetts Institute of Technology (MIT), aimed to explore theimpact intensive use of the I on brain activityfocusing on a precise task: the Writing of texts.
To do this, they were involved 54 peopleaged between 18 and 39, from five universities in the Boston area. Everyone was asked to write short essays wearing a kind of headphones equipped with electrodesable to record the electrical activity of the brain and to monitor the active areas and their connections during the various tests.
The participants were divided into three groups:
- Chatgpt Group: They could only use chatgpt as a source for writing, without accessing other resources.
- Google group: They could search freely online, but without using AI -based tools.
- BRAIN ONLY GROUP: They could not use neither external sources: they wrote only based on their knowledge.
Each participant wrote three essays in three different sessions. In one fourth sessionthe “chatgpt” and “brain” groups have exchanged methods. Whoever had used the AI before wrote without help, and vice versa.
And guess what was the result?
The brain activates less when we write with chatgpt
The participants who wrote the essays without any support outside showed the wider and intense brain connectivity, in particular in the areas associated with the processing of language,ofAttention focused and ideation Creative. The group that used Googleon the other hand, has activated more the regions involved in the search for informationin the’visual processing and in the memorythe areas we use when we navigate and summarize online sources. The group that has reached Chatgpt recorded the lower connectivity Among the various brain areas during the three tests. This, however,, does not mean What to use chatgpt “make it stupid“. The research team explicitly asked not to use expressions such as” brain rot “or” brain damage “. brain is less involved compared to autonomous writing.
An interesting fact emerges in the fourth session of the study: when the participants who had used chatgpts went on to write without any external help, theirs activity cerebral is increased Significantly, reaching levels similar to those of the “only brain” group. In other words: the brain reactivates when he returns to work independently. These results suggest that, before resorting to AI, it can be useful to try to solve problems alone. THE resultshowever interesting, still go interpreted with caution. Let’s see why.
Because we can’t say that Chatgpt makes us stupid
First of all, it is good to remind us that this article does not have yet past Peer-Reviewthe rigorous process in which other experts in the sector evaluate the quality and reliability of the study before the definitive publication. The MIT group has decided to pre-publish The study for raise the debate on an urgent theme, especially in view of possible political or school initiatives on theuse of AI by the youngest. But until there is an official review, the data must be interpreted with caution.
In addition to this, in the last month both the team itself that other experts have highlighted three critical issues Main:
Few people were involved. For this study only 54 people were involved, all belonging to the so -called group “Weird” (“Western, Educated, Industrialized, Rich, and Democratic”), or young adults from western and well -educated societies. This type of sample does not represent the entire world population and limits the generalization of the results.
The long -term results have not been evaluated. The experiment monitored the brain activity of the participants only during the writing sessions, distributed on a relatively short period. We do not know if and how the continuous use of chatgpt affects the brain over time, or if the effects observed are only temporary.
Writing from scratch and reviewing the text are not the same activity. Perhaps the most important criticism: those who write a text from scratch is engaged in a complete cognitive process – ideation, formulation, organization – while those who use chatgpt often carry out a work of revision, evaluation and modification of an already generated text. They are different mental activities. The lower brain activation observed in those who use AI could depend more on the type of task carried out than on an effective cognitive “drop”.
