Your Brain Deteriorates When You Use Generative AI To Replace Yourself Instead Of Complementing Yourself

With this technology, the seventh deadly sin becomes the first.

A study from | the Massachusetts Institute of Technology (MIT) warns about the risks of relying too heavily on generative artificial intelligence.

Researchers equipped 54 people with electroencephalography (EEG) headsets and asked them to write texts over a period of four months. One group used ChatGPT, another used Google Search, and a third relied solely on their own thinking.

By the end of the study, the participants who used ChatGPT showed significantly lower brain activity. When they later tried to write on their own, their brain function resembled that of beginners. Even more concerning, 83% of the ChatGPT users were unable to recall a single sentence from what they had written just minutes earlier—compared to only 11% among those who had used only their own minds.

In short, generative AI can simulate our brain—but it doesn’t stimulate it.

The MIT researchers coined a term to describe this effect: “cognitive debt.” It refers to the idea that we are borrowing against our future cognitive abilities in order to save time or reduce effort in the short term. MIT recommends an approach that matches how I personally use generative AI when writing important texts in English: don’t use it to write for you—use it to check what you’ve written.

In fact, the study showed that people who started a task on their own and finished it with ChatGPT had stronger brain activity than those who started with ChatGPT and completed the task on their own. As I explain in the generative AI training sessions I give, tools like ChatGPT should support—not replace—our intellectual effort. Like a good teacher, they are most helpful when they don’t give us the answer, but confirm the accuracy of the answer we’ve found ourselves.

The worst mistake with generative AI is, therefore, laziness.

Most of the MIT study participants had no idea their brains were being affected this way. It’s similar to the “Google Maps effect” – but applied to our thinking skills, not just our sense of direction. The same is true of how we increasingly outsource memory to our devices. The “extended mind” theory, introduced by philosophers David Chalmers and Andy Clark in 1998, suggests that, when we offload cognitive tasks to smartphones, those devices become part of our mental processes.

This isn’t new. David Chalmers and Andy Clark point out that the human brain evolved assuming we would use tools and interact with our environment – written language being a prime example. Reading isn’t hardwired into our DNA like speaking is. When children learn to read and write, the brain rewires itself: Visual pathways adapt to recognize letters and words, creating a specialized area for word processing. This physically reshapes the brain.

Humans have always looked for ways to reduce mental effort. But generative AI represents not just more of the same – it’s a shift in kind, not just in degree. Never before has a technology posed such a risk of degrading our cognitive function, while also giving us the misleading impression that we’re being smart. Used the wrong way, it can make us less intelligent.

To avoid piling up an unsustainable cognitive debt, we need to invest in the mental effort it takes to grow and protect our thinking skills – whatever they are.