A new study from the renowned MIT Media Lab has sparked considerable debate in recent days. Headlines are full of dramatic claims: using ChatGPT when writing supposedly makes the brain sluggish, weakens memory, and stifles creative thinking. Talk of AI causing “mental decline” spread rapidly. But on closer inspection, it becomes clear: while the study offers interesting insights, it’s far from proving that ChatGPT is inherently harmful to our minds.
The MIT researchers examined how using ChatGPT to help write essays affects brain activity. Participants were monitored using EEG scans. The result: those who used the AI showed reduced activity in certain parts of the brain – especially regions linked to creative thinking and complex problem solving – compared to those who wrote unaided. Many AI users could barely remember what they had written afterwards. That sounds alarming. But what’s really behind it?
First of all, the study focused on a very specific situation: 54 young adults from the Boston area were asked to write short texts under time pressure. Whether these findings apply to other types of AI use – such as coding, research, or casual communication – is entirely unclear. It’s also unknown whether people with more experience using AI tools would show the same effects.
And importantly, lower brain activity isn’t necessarily negative. It could equally be a sign of efficiency – perhaps because routine tasks require less mental effort with AI support. This doesn’t make us less intelligent; in the best case, it could even free up resources for other cognitive tasks. The researchers themselves urge caution when interpreting their data. They don’t see AI as inherently damaging, but warn that unthinking, long-term outsourcing of mental work could, over time, erode our own cognitive skills.
This highlights how crucial it is to consider how we use tools like ChatGPT. Those who use AI as an assistant – to develop ideas, structure thoughts, or refine writing – could actually benefit from it. But those who blindly rely on the tool and stop engaging their own minds may indeed risk weakening key abilities. And this is far from a new concern: similar debates arose with calculators, satnavs, or spellcheckers.
It’s also worth noting that this study is still a preprint – it hasn’t yet been peer-reviewed by independent experts. It offers thought-provoking early findings, but no definitive answers. The long-term effects of AI on our brains remain largely unstudied. In this field, science is only just beginning to scratch the surface.
The current headlines warning of the “dangerous AI” don’t tell the full story. They overlook the fact that, ultimately, it’s how we choose to engage with these tools that matters. ChatGPT isn’t a threat to our brains – unless we allow it to be by surrendering our thinking entirely. The MIT study offers a valuable reminder, not a condemnation of the technology itself. It reminds us that it’s not AI that shapes our thinking – but we ourselves.
