Misleading headline. This article is about ChatGPT and not "AI". A major problem with the "AI" grift is that it spawns so many dependent grifts like this one. "Science" under capitalism is not actually science.
post
The name and presentation of that site has a veneer of legitimacy, but it really doesn't seem credible.
There's also a lot of general antivax stuff.
Now, sharing a lot of ... Questionable articles... Doesn't make the article in question invalid. It does however call into extreme doubt any editorial context the site might be adding.
https://arxiv.org/pdf/2506.08872
This is the actual study being referenced. It's conclusions are significantly less severe than this presents them as, while still conveying "LLMs are not generally the best tool for facilitating education".
trade-off highlights an important educational concern: AI tools, while valuable for supporting performance, may unintentionally hinder deep cognitive processing, retention, and authentic engagement with written material. If users rely heavily on AI tools, they may achieve superficial fluency but fail to internalize the knowledge or feel a sense of ownership over it.
from an educational standpoint, these results suggest that strategic timing of AI tool introduction following initial self-driven effort may enhance engagement and neural integration. The corresponding EEG markers indicate this may be a more neurocognitively optimal sequence than consistent AI tool usage from the outset

Ultimately, this isn't saying AI tools cause brain damage or make you stupid. It's saying that learning via LLM often causes worse retention of the information being learned. It also says that search engines and LLMs can remove certain types of cognitive load that are not conducive to retention, making learning easier and faster in some cases where engagement can be kept high.
It's important to be clear and honest about what a study is saying, even if it's not as unequivocally negative as the venue might appreciate.
Just wanna point out that every time something scares you enough, it also reprograms/rewires your brain. Not trying to discredit the study, but the reprogramming really isn't the concern; it's if the reprogramming is beneficial, which this isn't.
The study appears to be saying something different from what the headline implies.
Basically it might be better to say that using an LLM doesn't require you to think as hard, you remember less of the essay, and when you go back to rewrite a previous essay without the LLM you have more trouble.
They also noted that for some people, using the LLM made them learn much better. Basically the difference between getting it to write for you and using it as a tool to structure information.
One reduced cognitive load from all sources, and the other reduced load relating to integrating different information sources.
Basically it was a proper study by people who knew what they were doing. They never actually said anything about rewiring.
This comment just reprogrammed my brain after the reprogramming it got from that post 😮
using AI images in an article about AI use leading to cognitive decline gotta be crazy💀
Yeah but what do you expect them to do, actually pay a human to make sure they don't do that?
Or just survive on the merit of the text content?
Does context escape your brain, The images are not the focus of this article, the fucking article is you weirdo
Amusingly using what appears to be an AI generated image.
Tech bros stay losing, it is a good day
Hooray!
We invented cyberpsychosis, for reals!
Isn't it so cool to live in a cyberpunk dystopia!?!
Brb, gonna go OD on some early access/preview alpha braindances!

Why would she need a VPN to access chat gpt though?
There’s a good body of research on cognitive capacity and creativity in regard to enriching environments. Even down to rats. Give rats playgrounds and toys and they perform better at memory tasks and solving puzzles.
I suppose you could train rats to press a button to get a human to come solve problems for them. Take the human away, then what?
What’s insidious here is the same over scheduled kids, having their childhoods choreographed for enrichment, are often the ones coming out of childhood with critical thinking, anxiety, and socialization deficits, we think because they’re using their phones for every problem solve.
Related post from when this was initially published: https://midwest.social/post/30021473
The study itself is entirely benign, and I'd actually accept it as a reason to eschew AI in an educational context. Their conclusion is basically "if you use an LLM to write an essay you tend to not retain the information as well", which is... Downright boring in how reasonable it is. Particularly given the converse observation I wouldn't have expected: if you are already familiar with a subject then using an LLM to write an essay can strengthen your understanding.
The "journal" this summary of the study was shared in is quackery, so I'm not surprised they distorted the findings.
Nothing bad ever happens.
No, like, you're seemingly saying "new inventions never cause health problems"
Y'know, like asbestos. It was a wonder material! And then the health effects were uncovered.
Literally every piece of technology introduced in the past thousand years has had this kind of hue and cry built up around it, beginning with the printing press and books in Europe.
The ill-concieved printing press argument is a standard pro-AI trope.
Remember when television would "rot kids' minds"? Remember when the Internet was going to end civilization as we know it?
Yeah, i don't know where you live but here in America our democracy is in shreds thanks to those things. They came true, just as they said it would. Environmental collapse is next, for those keeping score at home.
Saying knee jerk rejection of new technology is common is pro-ai?
If it's such a bad, ill-concieved notion, why don't you explain why it's wrong, instead of just saying that it's used by people you disagree with?
Maybe if an argument is used by both pro and anti AI people, it's less a "pro-ai" argument, and more a "let's keep in mind how often doom and gloom has been wrong and keep our criticism grounded"?
If it's such a bad, ill-concieved notion, why don't you explain why it's wrong, instead of just saying that it's used by people you disagree with?
Again? No. Maybe someone else feels like explaining it this time.
"let's keep in mind how often doom and gloom has been wrong and keep our criticism grounded"?
My good dude, the sentient life on this planet is about to witness several extinction events in our lifetime because we ignored the grounded "doom and gloom" research. If you want to stand on the big red X this time while a loud whistling noise and billionaire cackling gets louder, that's up to you but as many people have tried to say before - it's a bad idea, don't do it.
Oh, baby, did you read the thing? Something tells me you didn't read the thing.
Did they control this against my mental decline after using a rage-inducing traditional search for 10x as long? I get the point being made, but nothing exists in a vacuum.
Actually, yes. I know you were making a point, but the study this pile of garbage article referenced actually did comparisons between people using an LLM, people using a search engine, and people using neither.
The study is a lot more reasonable. It basically says using an LLM to write a research paper on a subject makes you retain less of the subject matter.
They specifically mention the cognitive load of using the tool, be it search engine or LLM, and how that load doesn't contribute to knowledge retention.
Their research indicates "over usage" as opposed to "any usage" is bad for learning a subject.
Some people think doing work makes them weaker at the end of the day. Other people do work and feel it makes them stronger for having done it so you’re clearly in the first cohort.

Living the cosmic horror, a god we didn't know we were summoning is changing our minds for the dumber.
Maybe the Amish were on to something.
I know this is true because I've seen people get dumber using their little bots for everything. But techbro will argue that you could say the exact same thing about the internet or electricity. Technically all inventions have made us dumber. The majority of people cant make clothes, do long division, build a log cabin etc because its all been made easier.
They could do those things (at least some of them) if they were taught and tried to do them, though. Someone else making your clothes for you is older than civilization, it’s basic task specialization. I think giving up thinking altogether is a different problem than giving up knowing how to make every single tool that you need to survive.
Ironic that the article is using shitty AI generated images to illustrate the point.
i gotta say, not using LLMs is more fun than using them, but I do have the luxury of time.
1 comment