No more tossing and turning at midnight panicking about that history paper you haven’t even started. No more staring at a blank screen trying to come up with the perfect opening sentence for your elevator pitch. Since 2022, Artificial Intelligence (AI) models, such as ChatGPT and Google Gemini, have increasingly been taking over tasks in our everyday lives. But where does the extensive use of AI leave our brain? Aren’t we getting dumber while AI is getting smarter?

A new best friend

ChatGPT and Google Gemini are examples of ‘Large Language Models’ (LLMs). An LMM is a type of Artificial Intelligence that can interpret and generate human language. A LLM learns patterns and rules of language, similar to the way a human learns to communicate through exposure to language. After training, it can help you with things like answering questions, summarising texts, translating, and writing content.

In education, students turn to AI to find quick answers, to automate boring tasks, or to personalise their learning experiences. Not only can it help students do work more quickly, it can do much of the work for them. This convenience, however, also raises concerns. Some researchers argue that our overall thinking abilities, our brain power, may be weakening. Could this be happening because we are handing over more of our thinking to AI? If AI starts doing too much of the thinking, could it ultimately make us less intelligent?

Our IQ is decreasing

One argument that we may be becoming less intelligent is linked to the ‘Flynn effect’. The social scientist James Flynn determined that the average IQ  steadily increases by about 3 points per decade. But in recent decades, this effect has slowed or even reversed. This reversed effect has also been observed in the Netherlands. The question now is, of course, what is causing this?  Since intelligence is shaped by many things, there is no simple answer. Still, spending much time on technology (including the use of AI) could be one piece of the puzzle behind this decline.

Is AI really making us dumber?

So far, there is no proof that AI is making us dumber. However, priliminary research suggests that if we let AI do too much of our thinking, we may train our brains to put in less effort and think less deeply. One study looked into this by letting people write several essays. Fifty-four people were divided into one of three groups. The first group was asked to write an essay using ChatGPT (LLM); the second could only use the internet (search engine), but no AI; and the third could not use any external information sources at all (brain-only). While they wrote, researchers measured their brain activity. This revealed significant differences in the brain activity between the three different groups.

The ‘brain-only’ group was the most engaged with their task. They had to actively think about the assignment, which means they were actually learning. They also claimed feeling more curious and having a stronger sense of ownership over their essays. Their effort was thus rewarded with a positive feeling, comparable for example to the satisfying feeling after a good workout. When later asked to rewrite their essays from memory, they did so successfully. The people who used the internet also demonstrated relatively strong brain activity and high satisfaction.

In contrast, the LLM users showed the weakest brain connectivity. When it came time to rewrite their essays from memory, many were not succesfull since the material had not really stuck. This leads to a simple, but important conclusion: learning sticks better when you actively search for and process information. And that is the  point: struggling with a task isn’t a flaw. It is how we build understanding, critical thinking skills, and long-term memory. Of course, this is one small study focused on a specific task, but it illustrates how outsourcing thinking to AI can change what happens in our brains while we work.

AI is like a calculator

Admittedly, concerns about new technology are nothing new. Over the past years, generative AI has often been compared to the handheld calculator. When handheld calculators were first introduced, many people feared that students would stop learning basic math. In reality, calculators became helpful tools that allowed students to solve even more advanced problems instead. Some people argue that AI could play a similar role. They argue that if used properly, AI can increase productivity and free up time for deeper thinking.

Many researchers, however, criticise this comparison. Calculators still require you to understand the maths. They speed up part of the process but don’t solve the whole problem for you. AI, in contrast, can generate the entire answer, sometimes without you understanding the steps at all. That is exactly where the risk lies.

The real power of AI lies in using it with care, so it can work like a calculator. For example, using it to review your essay instead of letting it write the essay for you. It works best when paired with human thinking, not when used as a replacement. AI is good for handling fast, repetitive tasks and generating drafts, while humans remain stronger in creativity, ethical reasoning, emotional understanding and deep critical thinking. Balance is the key: AI should amplify what you can do, not take over the vital work that high-quality learning requires.

This article was written by Rinke van der MeulenExternal link

Sources

Dutton, E., Van der Linden, D., & Lynn, R. (2016). The negative Flynn Effect: A systematic literature review. Intelligence, 59, 163–169. https://doi.org/10.1016/j.intell.2016.10.002External link

Kosmyna, N., Hauptmann, E., Yuan, Y. T., Situ, J., Liao, X.-H., Beresnitzky, A. V., Braunstein, I., & Maes, P. (2025). Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task. arXiv. https://doi.org/10.48550/arXiv.2506.08872External link