(Source: Chaikom/Shutterstock)
It’s been theorized that using AI tools like ChatGPT to help with writing tasks will lead to lower brain activity. Now, researchers at the MIT Media Lab have discovered strong evidence of that in a newly published study.
In the recently published study, titled “Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task,” MIT Media Lab researchers sought to determine the impact of AI on brain activity during writing exercises.
The study, which you can download here, consisted of three cohorts: one group that used only their brain, one that used a search engine, and one that used a large language model (LLM) like ChatGPT.
The measurements consisted of electroencephalography (EEG) to measure participants’ brain activity and to see how their neurons activated during the essay writing task. Researchers also performed NLP (natural language processing) analyses on the essays, including the use of English teachers and an AI judge. They also interviewed the subjects after each session, in which they scored the participants on their recall.
The researchers moved the participants among the different groups across four sessions to determine what impact the different writing modes would have on their brain activity, NLP, and recall tests. The researchers discovered that the different groups had “significantly different neural connectivity patterns, reflecting divergent cognitive strategies.”
Critically, the use of AI tools correlated with poorer results, both in brain activity and the other tests. The researchers write:
Brain activity decreases as AI use increases. (Source: “Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task”)
“Brain connectivity systematically scaled down with the amount of external support: the Brain‑only group exhibited the strongest, widest‑ranging networks, Search Engine group showed intermediate engagement, and LLM assistance elicited the weakest overall coupling.”
Study subjects who went from the LLM to the brain-only group showed weaker neural connectivity and “under-engagement of alpha and beta networks,” the researchers write. Subjects who went the other way, from the brain-only group to the LLM group, showed higher memory recall and re-engagement of activity in certain parts of the brain.
Using ChatGPT to write your essay? Your English teacher is not impressed. (Source: Net Vector/Shutterstock)
The search engine group fell in the middle in these tests, with stronger results than the LLM group but weaker than the brain-only group. Members of the LLM group, the researchers note, scored poorly when asked to quote from an essay they had “written” just minutes earlier.
The English teachers reported that it was obvious which essays were written by AI because they were “near-perfect” in their use of language and structure, and for how devoid they were of any personal touch. “We, as English teachers, perceived these essays as ‘soulless’, in a way, as many sentences were empty with regard to content and essays lacked personal nuances,” they wrote, according to the study.
Overall, the results became clear: Using AI tools, such as LLMs, has a negative impact on brain activity.
“The use of LLM had a measurable impact on participants, and while the benefits were initially apparent, as we demonstrated over the course of 4 months, the LLM group’s participants performed worse than their counterparts in the Brain-only group at all levels: neural, linguistic, scoring,” the researchers write.
Our society may be on the path to artificial general intelligence (AGI) or super-intelligence. These artificial neural networks may help us solve great scientific and engineering challenges. But in the meantime, relying on AI tools may be harming the brain functions of individuals.
This article first appeared on our sister publication BigDATAwire.
Related
