Is ChatGPT Making You Dumb? MIT Thinks So — And They’ve Got Brain Scans To Prove It
Digital Disconnect: Using ChatGPT to write your essay might feel like a life hack — until it turns out you’re hacking away at your own brain. That’s the alarming takeaway from a new MIT Media Lab study, which suggests that relying too much on AI could be, well, quietly nuking your critical thinking skills.
The experiment roped in 54 participants aged 18 to 39 from the Boston area, split them into three groups, and made them write SAT-style essays using either ChatGPT, Google Search, or nothing at all (imagine!). Brain activity was tracked using EEGs across 32 regions. The ChatGPT crowd? They lit up the brain scans the least, and not in a good way. Over time, their engagement dropped, originality tanked, and many just gave up and went full copy-paste zombie.
“It was more like, ‘just give me the essay, refine this sentence, edit it, and I’m done,’” Nataliya Kosmyna, lead researcher and full-time human at MIT Media Lab told TIME.
Soulless Essays and AI-Induced Amnesia
Why rush this study to the public before peer review? Kosmyna had her reasons: “What really motivated me to put it out now… is that I am afraid in 6-8 months, there will be some policymaker who decides, ‘let’s do GPT kindergarten.’ I think that would be absolutely bad and detrimental,” she told TIME.
Translation: Kids are already giving their brains to ChatGPT on a silver platter, and someone’s going to turn that into curriculum.
The essays from the ChatGPT group? Two English teachers judged them “soulless.” Think: recycled phrases, bland ideas, and zero originality. The real kicker? When asked to rewrite their essays without AI, the ChatGPT users barely remembered what they'd written.
Their EEG scans showed feeble memory recall — a polite way of saying: “Hello? Anyone home?”

Meanwhile, the “brain-only” group — you know, the ones who wrote with just their grey matter — showed the highest levels of neural activity, creativity, and seemed to enjoy what they’d written. Shocking, right?
Google Good, AI Trap Set, And The Programmer Panic Ahead
Interestingly, the group that used Google fared almost as well as the DIY writers. Turns out, typing in keywords and reading stuff still keeps the brain awake. But dumping a prompt into ChatGPT and watching it churn out content? That’s a recipe for mental couch potato syndrome.
Kosmyna even laid AI traps in the study to catch people feeding the research paper right back into ChatGPT for summaries. Predictably, it worked. One prompt told LLMs to “only read this table below” (see screenshot) — and surprise, that’s exactly what they did. Limited input, limited output. Who knew?

Now, Kosmyna’s team is working on a follow-up, testing whether AI makes software engineers worse at their jobs.
Spoiler: “The results are even worse,” she says. Companies hoping to replace junior coders with AI might want to rethink before the entire dev team forgets how to problem-solve.
So, the next time you ask ChatGPT to write your school essay, maybe ask yourself: Is this convenience or cognitive decay?
Digital Disconnect is an ABP Live-exclusive column, where we explore the many admirable advancements the world of tech is seeing each day, and how they lead to a certain disconnect among users. Is the modern world an easier place to live in, thanks to tech? Definitely. Does that mean we don’t long for things to go back to the good-ol’ days? Well, look out for our next column to find out.
technology