In the midst of all the debates on how AI affects jobs, science, the environment and everything else, there is a question of how large language models have an impact directly on people who use them directly.
A new study of MIT Media Lab implies that the use of AI tools reduces brain activity in certain respects, which is naturally alarming. But I think it is only part of the story. The way we use AI, like any other element of technology, that’s what really matters.
Here is what the researchers have done to test the effect of AI on the brain: they asked 54 students to write trials using one of the three methods: their own brain, a search engine or an AI assistant, specifically chatgpt.
On three sessions, students stayed with their assigned tools. Then they exchanged, with tool users without tools and non-tool users using AI.
The EEG helmets have measured their brain activity throughout, and a group of humans, as well as a specially trained AI, have marked the resulting tests. The researchers also interviewed each student of their experience.
As you can expect, the group counting on their brain has shown the most commitment, the best memory and the most sense of the property of their work, as evidenced by how much they could quote.
Those using AI at first had less impressive recall and brain connectivity, and could not even cite their own trials after a few minutes. During writing manually in the final test, they have always underperformed.
The authors ensure that the study has not yet been evaluated by peers. It was limited in the scope, focused on writing tests, not another cognitive activity. And the EEG, although fascinating, is better to the extent of global trends than to pinch the exact brain functions. Despite all these warnings, the message that most people would withdraw is that using AI could make you stupid.
But I reappeared this to consider if perhaps that AI does not stupid us as much as to let ourselves withdraw from the reflection. Maybe the problem is not the tool, but how we use it.
IA brain
If you use AI, think about how you have used it. Have you written it a letter, or perhaps think about some ideas? Has this replaced your reflection or did it support it? There is a huge difference between the outsourcing of a test and the use of an AI to help organize a disorderly idea.
Part of the question is that “AI” as we refer it is not literally intelligent, just a very sophisticated parrot with a huge library in his memory. But this study did not ask participants to think about this distinction.
The Using LLM group was encouraged to use AI as they considered it good, which probably did not mean thoughtful and judicious use, it is enough to copy without reading, and that is why the context is important.
Because the “cognitive cost” of the AI can be less linked to its presence and more to its objective. If I use the AI to rewrite an e-mail bridge, I do not decrease my intelligence. Instead, I release a bandwidth for things that really require my reflection and my creativity, such as offering this idea for an article or plan my weekend.
Of course, if I use AI to generate ideas that I never take the trouble to understand or commit, then my brain probably takes a nap, but if I use it to rationalize the tedious tasks, I have more brain when it is important.
Think about it like that. When I grew up, I had dozens of phone numbers, addresses, birthdays and other details of my friends and family memorized. I had written the majority somewhere, but I rarely needed to consult it for those I am closest. But I haven’t memorized a number for almost a decade.
I don’t even know my own fixed number by heart. Is it a sign that I become more stupid, or just evidence that I had a mobile phone for a long time and that I stopped needing to remember it?
We have unloaded certain types of reminders from our devices, which allows us to focus on different types of thought. Competence does not memorize is how to find, filter and apply information when we need it. It is sometimes called “Extelligence”, but it is really a question of applying the power of the brain where it is necessary.
This does not mean that memory no longer matters. But the accent has changed. Just as we do not only make the students practicing a long division by hand once they understand the concept, we can one day decide that it is more important to know what a good form of form looks like and how to encourage an AI to write one than to write it online from zero.
Humans always redefine intelligence. There are many ways to be intelligent, and knowing how to use tools and technology is an important measure of intelligence. At one point, being intelligent meant how Knap Flint, make Latin variations or work a slide rule.
Today, this could mean being able to collaborate with machines without letting them make all the reflection for you. Different tools prioritize different cognitive skills. And whenever a new tool arrives, some people panic that it will ruin or replace us.
The press to print. The calculator. Internet. All were accused of having made people of lazy thinkers. Everything turned out to be a big boon for civilization (well, the jury is still on the internet).
With the AI in the mixture, we are probably looking stronger about synthesis, discernment and emotional intelligence – the human parts of the human being. We don’t need the kind of scribes that are not good for writing what people say; We need people who know how to ask better questions.
Know when to trust a model and when returns. This means transforming a tool capable of doing the work into an asset that helps you do it better.
But none of this works if you treat AI as an automatic distributor for intelligence. Punch in an prompt, wait for the brilliant fall? No, that’s not how it works. And if that’s all you do with it, you don’t become more stupid, you have never learned to stay in touch with your own thoughts.
In the study, the possession of writing of the LLM group did not only concern memory. It was a question of commitment. They did not feel connected to what they wrote because they were not those who were writing. It is not a question of AI. It’s about using a tool to ignore the difficult part, which means skipping learning.
The study is however important. This reminds us that the tools shape thought. This pushes us if we use AI tools to extend our brain or to avoid using them. But to claim that the use of AI makes people less intelligent is like saying that the calculators hurt me in mathematics. If we want to keep our brain bright, maybe the answer may not be to avoid AI but to think about using it.
The future is not the human brain against AI. These are humans who know how to think with AI and any other tool, and avoid becoming someone who does not bother to think at all. And it’s a test that I would always like to go through.
