New research indicates that ChatGPT may negatively impact individuals’ critical thinking skills over time.
At MIT’s Media Lab, researchers conducted a study where participants were divided into three groups. One group used OpenAI’s ChatGPT, another utilized Google’s search engine, and the third didn’t use any tools, referred to as the “brain-only” group. Throughout the study, participants’ brain activities were recorded using EEG to monitor different brain regions.
The findings revealed that participants who interacted with ChatGPT for several months exhibited the lowest levels of brain engagement, displaying “consistent reductions in neural, verbal, and behavioral performance,” as noted in the survey.
Initially, users of ChatGPT employed the large language model to pose structural questions about their essays. However, by the study’s end, many were more inclined to copy and paste existing content.
In contrast, those using Google’s search engine showed moderate brain activity, while the “brain-only” group demonstrated robust network engagement.
These results suggest that reliance on large language models could deteriorate cognitive abilities over time, particularly among younger users, raising concerns as AI becomes increasingly prevalent in education.
Natalya Kosmina, the lead author, expressed urgency in releasing the findings before a complete peer review due to the potential for policymakers to jump into implementing “GPT kindergarten” in schools within months. She believes this could be detrimental to brain development, describing the risk as significant.
Meanwhile, the integration of AI in education is not waning. In April, an executive order was signed by President Trump aiming to embed AI into American classrooms.
The executive order’s primary goal is to prepare students for the future workforce by ensuring they are adequately trained with AI tools.
