A new study conducted by MIT’s Media Lab found that ChatGPT harms one’s critical thinking abilities. While the study has not yet been peer-reviewed and its sample size is small – they surveyed 54 subjects between the ages of 18 to 38 years old – they discovered that those asked to use ChatGPT to write SAT essays (the first group) had the lowest brain engagement and “consistently underperformed at neural, linguistic, and behavioral levels.” From observing this group over a number of months, the researchers found that ChatGPT users got lazier with each essay. 

The study’s primary author, Nataliya Kosmyna, and her colleagues found that the group they asked to use ChatGPT wrote similar essays that lacked original thought, using the same expressions and ideas. When two English teachers reviewed the essays, they remarked that they were “soulless”. The second group, known as the brain-only group, had the highest neural connectivity, especially in alpha, theta and delta bands, which are associated with creativity ideation, memory load and semantic processing. According to the researchers, this group was more engaged and curious and claimed ownership over their work, expressing higher satisfaction with their essays. The third group was allowed to use Google Search to support their work, and they expressed high levels of satisfaction and active brain function. 

After writing their essays, the subjects were then asked to rewrite one of them, but the ChatGPT group had to do so without the tool. The brain-only group could use ChatGPT. In this section of the study, they found that the ChatGPT group remembered very little of their original essays, showcasing weaker alpha and theta brain waves, which demonstrated a bypassing of deep memory processing. The brain-only group performed exceptionally well at this task, showcasing a significant increase in brain connectivity across all EEG frequency bands. This suggests that perhaps ChatGPT could help enhance learning if used properly. 

As aforementioned, this study has not yet been peer-reviewed. Kosmyna and her team did not want to wait for approval before sharing the results, as peer reviews can take up to eight or more months. Kosmyna and her team are particularly worried about the impact generative artificial intelligence chatbots have on young people’s cognitive development. “Education on how we use these tools, and promoting the fact that your brain does need to develop in a more analogue way, is absolutely critical,” Kosmyna stresses to Time. “We need to have active legislation in sync and, more importantly, be testing these tools before we implement them.”