Can we keep thinking only about the future? According to a study by MIT, a prestigious American university, AI runs the risk of giving a severe blow to “critical thinking.” In any case, these are the results that emanate from an experience carried out in 54 people aged 18 to 39, having to write a dissertation.
A first group could use Chatgpt, a Google second and a third of any search engine. All were equipped with sensors to register their brain activity, so that the group used by Chatgpt had a “weaker” brain commitment, says time, which the study reveals.
A brain that no longer wants to find for itself
“The subjects of the low performance experience constantly at the neuronal, linguistic and behavioral level,” explains Nataliya Kosmyna, the main author of the study. He has also decided to communicate about his work, while everyone is passionate about generative artificial intelligence.
Alerts, in the face of these results, about the risk of allowing children, or even very young children, use AI: “I think it would be absolutely bad and harmful, development brains are at greater risk,” he explains.
Another MIT study, published in early 2025, had established that a user who spends a lot of time chatting with Chatgpt had more feeling of being alone.
For this new experience, it seems that people who use Chatgpt have delivered a relatively similar text, based on the same expressions and the same ideas, without any original element. Like the experience led by Tech & Co on the subjects of the BAC philosophy of BAC, the concealers considered the result “without a soul.”
On the contrary, the other two groups that use Google or who have to administer for themselves have shown more satisfaction and participation in the work to be carried out.
Young people who have become addicted to AI
In the rest of the experience, while the researchers asked the three groups to rewrite one of their text, and by prohibiting the IA Use Chatgpt group, the latter did not remember what he had been able to write. On the contrary, in the group that had no right to anything, they were allowed to help with chatgpt, and the results were more convincing.
For the MIT, the demonstration is important: it is used to demonstrate that AI should not be used anyway. Therefore, you must improve learning and not decrease it by delivering a ready -to -use result.
Questioned by the Times, psychiatrist Zishan Khan is alarmed on his part of the constant propensity of adolescents based strongly on AI for their duties: “There is an excessive dependence on the great models of language, and this can have involuntary psychological and cognitive consequences.”
Keep in mind that the study now disseminated has not yet been read by the classmates and, therefore, must still be taken in retrospect.
MIT provides other studies of this type. The following should make it possible to shed light on the consequences of generative AI in the field of programming.
Source: BFM TV
