The ChatGPT group “consistently underperformed at neural, linguistic, and behavioral levels..”
Researchers at MIT’s Media Lab have raised alarm over ChatGPT’s impact on users’ brains. In a study pending peer review, scientists scanned the brains of 54 adults using electroencephalogram (EEG) while writing essays. Participants were divided into three groups: those using ChatGPT, Google Search, and no AI tools.
The ChatGPT group “consistently underperformed at neural, linguistic, and behavioral levels” and showed “weaker neural connectivity and under-engagement of alpha and beta networks.” In contrast, the group using no AI exhibited the strongest cognitive activity.
Prior studies have echoed similar concerns. MIT earlier found signs of “addiction” and “withdrawal symptoms” among ChatGPT “power users.” A Carnegie Mellon and Microsoft study noted reduced critical thinking.
Mental health concerns are rising too. Some users reportedly grew obsessed and ignored psychiatric medication after chatbot advice.
OpenAI responded: “We know people use ChatGPT in deeply personal moments… We’ve built in safeguards to reduce the chance it reinforces harmful ideas.”
The study adds to growing worries over AI’s cognitive and psychological effects.