Why I’m Afraid of What ChatGPT Is Doing to My Brain
A good friend of mine sent me an MIT article that hit me harder than expected.
What if AI is slowly making us… dumber?
I was a bit concerned by the headlines like “With ChatGPT, neural connections collapsed by 47%!” “Users can’t remember what they wrote!”
Nothing new, though. We’ve heard this before: Screens are bad. We scroll too much. Dopamine is the new sugar. But something felt different this time.
With the rise of ChatGPT and other large language models (LLMs), I’m really afraid of the impact on my cognitive competencies and social skills.
🧠 We’re Outsourcing Our Thinking
Humans evolved by thinking, imagining, and struggling. That’s our animal realm superpower; that’s how Homo Sapiens got here. But now, I see more and more people (myself included, 200%!) slowly letting go of that gift.
When I wrote my thesis on innovation in the luxury industry, it took me over 50 hours of deep work. Reading long and boring books. Thinking. Struggling to write.
Today, I could write the same thesis in 10 minutes with ChatGPT. Yes, I’d save time. But what am I really doing with the 49 hours saved? Scrolling Reels on Instagram? Binge-watching Netflix? Drinking wine with friends? Swiping on Tinder?
That time doesn’t make me smarter. If anything, it melts my brain further.
⚠️ What Science Actually Says
Today, we have only short-term and small scientific experiments at our disposal. There is no long-term evidence that LLMs destroy brains. But the first MIT study about this subject give a powerful early signal.
- 54 people wrote essays using either:
- Only their brain
- Google Search
- ChatGPT
Using brain scans, they discovered that ChatGPT users had the lowest neural activity. Their essays were quite repetitive and soulless.
Even worse, when asked to rewrite their own essays without ChatGPT, they couldn’t recall what they had written. At the same time, the “brain-only” group showed the highest levels of attention, creativity, and ownership.
😶🌫️ Loneliness & LLMs: Another Red Flag
Another red flag for me is the emotional impact of AI. So let’s hypothetically say that ChatGPT is dulling our minds. Additionally, it makes us feel more lonely, too.
MIT and OpenAI joint research found that the more people talked to ChatGPT, the lonelier they felt. Some users formed emotional attachments to the chatbot. Others felt disconnected from real human interaction.
Personally, I can converse with ChatGPT for long minutes, but I’ve become extremely bored with small talk.
Compared with ChatGPT, where I can have very long and deep conversations, the usual chitchat in bars… I’ve become so bad at it.
The pattern was clear: overuse of AI = less socialization + more loneliness.
I’m Letting Tech Think for Myself.
- Mental math? Replaced by calculators.
- Remembering things? Google can do it better.
- Writing, summarizing, analyzing? Now ChatGPT and its cousins do it better and faster.
The more I let go of those little cognitive tasks, the more U forget how to do them.
I’m totally in for those transitions. However, I’m cautious, and I notice the effect on my brain. This statement becomes even more worrying when we look at the younger generation:
“From a psychiatric standpoint, I see that overreliance on these LLMs can have unintended psychological and cognitive consequences, especially for young people whose brains are still developing. These neural connections that help you in accessing information, the memory of facts, and the ability to be resilient all that is going to weaken.”
— Psychiatrist Dr. Zishan Khan in the Times
The Real Danger: It’s Not Just You and Me, It’s Society
I’m scared we’re becoming dumb, and that it’s becoming visible… even in democratic elections.
Because less thinking means fewer ideas. Less attention means more vulnerability to disinformation. And when those patterns scale up to society, we get absurd public debates and poor decision-making even at the top levels.
We can’t afford that.
Another Ecological Absurdity
Something else we shouldn’t afford. There’s also an ecological side to all of this. We all know how energy-intensive prompting is.
The massive carbon footprint, the tons of water used to cool the servers, and the huge electricity bills all spent to train these AI models.
The truth is: the biggest part of the environmental cost is already behind us, most of which was used during training. The models exist. High tier models exist. Midjourney exists. Suno exists. The tools are here.
Now, we will use those tools anyway. There’s no logic in going back. It’s like digging up uranium, building a nuclear plant… and then refusing to turn on the power.
So now, the least we can do is use them meaningfully.
- Not to write faster Slack messages or punctual emails.
- Not to generate yet another “5 tips to be productive” carousel.
But to create something that expands our human potential.
I may not agree with everything my ecological mentors Jean-Marc Jancovici or Aurélien Barreau say on this topic especially when they advocate pausing AI altogether. But I do think: we don’t need GPT-5 to make progress. The UI of some 2023 LLMs is already a small miracle.
Let’s stop chasing the next model and start using what we have well.
So What Now?
We’re all guilty of the same screen addiction. And our managers will expect us to be more productive with these GenAI and LLM tools and to produce more for the same salary.
So resistance isn’t going to be easy…
We all agree on the fact: “we want to read more, scroll less, produce more, talk more, prompt less.”
And then… we don’t do it.
To write this article, I had to resist the urge to check my messages (on WhatsApp, Instagram, Messenger, etc.). I had to stop myself from prompting every paragraph for syntax mistakes.
If you’re reading this without automatic summarizing, and you’ve made it all the way here without checking your phone, like me, you’re already part of the cognitive resistance.
We will likely feel lonely during this revolution. Tech-woke, weirdo, unefficient, unproductive, stupid and other nice compliments. But don’t forget that we do it for our own sanity.
So… What Do We Do?
I honestly have no idea. And that’s why I’m writing to help clear my mind.
A few ideas though:
1. Digital detox (always good in general)
Reclaim your attention.
- Set screen-free hours (I use Opal on iOS).
- Uninstall apps (I’ve deleted Instagram at least 125 times).
- No phone during meals or Netflix.
- Boredom is a superpower. Rediscover it.
2. Train your brain without LLMs
Flex that mental muscle.
- Do quick mental math.
- Write on your own.
- Solve puzzles, logic games, and board games.
- Have real conversations with nuance, arguments, and maybe disagreement.
- Don’t outsource everything to ChatGPT or other LLMs.
3. Read
Not IG captions or Reels subtitles, let’s read real books. Without notifications.
For many specialists, it’s survival training for our brains (Source 1, Source 2, Source 3 ).
If this article made you pause, even just for 2 minutes, do yourself a favor:
Close this tab.
Put down your phone.
Grab a book. Or go for a walk.
Your brain will thank you later.
