Study: ChatGPT is Bad for Your Brain

Post Reply New Topic

Cognitive activity scales down with what the study describes as tool use, which is to say outsourcing all the hard work to an LLM gradually makes you stupid.

too premature to come to this conclusion. Chatgpt is just an advanced form of google search. People still need to read and verify sources.

Any time you don't engage in an activity, the body allows those parts of the body to atrophy. Don't use your muscles and sit around all day? You lose strength. Don't use parts of your brain? They weaken. Whether it's bone-density or brain activity, if you don't use it, you lose it. You can get it back, but it takes more work to get it back than it did to keep it in the first place.

The same patterns have been found in people who engage in passive entertainment (doomscrolling, binge-watching, media consumption, anything where your brain just passively absorbs stuff) vs people who engage in active hobbies, like reading actual books, building models, learning an instrument, things which actually activate the brain.

Certain people don't even verify their sources even when they don't use chatgpt, so why would they start checking them now, with chatgpt? Anyways, chatgpt is a lot more than just an advanced google search. That's why we have a topic about how colleges are cracking down on chatgpt usage as relates to cheating and plagiarism. Ghatgpt can create entire papers, just based on some prompts, with minimal user interaction.

An AI like Grok can even tailor writing to specific styles. Perhaps chatgpt can too, I've not tried. But basically, I could tell it to analyze the collective works of earnest hemmingway, and write a persuasive paper comparing hemmingway's outlook on life with the philosophies of emmanuael kant, based on any similarities stated or implied, to format it according to APA7 guidelines, include enough detail to make it at least 7 pages in length, and to write it in the style of william shakespear - and it would do it's best to spit out exactly that - with surprising effectiveness.

Actually brain rot https://en.wikipedia.org/wiki/Brain_rot as endorsed by Oxford Uni Press https://corp.oup.com/word-of-the-year/, already started at the beginning of the internet era long before chatgpt.

Anyone comes across anything they didn't understand, professor google (or any search engine) has always provided instant answers since the early 2000s.

this is an entirely different matter as it pertains to demonstrating the application of higher order thinking as per "Bloom's taxonomy pyramid" or in other words "mental gymnastics". the average person outside of university or professional life has no functional purpose in writing essays or reports.

Competency and capacity in using output from LLMs to produce papers is evolving and we are simply in a transition phase not unlike the era during the 20th century when students started using calculators to replace long division and computers replacing handwritten work. I remember the same conversations about the death of education and young people becoming dumber during that era. All that happened is we all evolved.

Since ancient times stories have been a vehicle by which people learned. even an essay or lab report is basically a story in technical form. the way we engage with stories has changed. A simple example is many people prefer a video rendition of a book in the form of a movie or documentary. In the youtube era you can get 20min edited movie summaries. Point is, people will still be curious and want to experience Hemmingway or watch a documentary of Kant's work. Also we are stuck in 2025 mode in thinking this is the end. Hardly. in as little as 10 years we can probably download the entire works of every writer and every philosopher into our long term memories via neural chips.

Our great grand children might well be the first generation of cyborgs able to access/download the entire knowledge base of the planet. the real interesting question will be how knowledge acquisition will intersect with developmental psychology in several decades?

So, which is it? It's too early to conclusively conclude brain rot is a real consequence of outsourcing one's thinking to AI, or brainrot is a widely documented phenomenon?

You seem to espouse a lot of Schrodinger's opinions.

too early...anyway I am somewhat doubtful chatgpt will make much difference to our collective brains, the bar has already been set very low thanks to social media/google and the overwhelming need for bite sized online content, this all happened before AI. LLMs can't really make the collective cognitive lazyness much worse.

I bet if you do an analysis of average post word counts from the time I joined wrong Planet in 2011-2025 I guarantee it's significantly decreased. Reading capacity thresholds have lowered and patience to write has diminished. All that has nothing to do with chatgpt.

Last edited by cyberdora on 07 Jul 2025, 12:11 am, edited 1 time in total.

I mean, if the use of google is supposedly already responsible for brain-rot, then logically, wouldn't "advanced google" cause even more advanced brain rot?

The cool thing about googling things you don't understand, is that now you can regurgitate a bunch of words you saw, while still not understanding it.

I'm still chuckling at hand-waving away the utility of "Bloom's taxonomy pyramid" (as well as calling it that), and comparing it to "mental gymnastics".

You say LLMs can't make intellectual laziness much worse, despite them allowing for laziness of a whole order of magnitude greater because they'll pretend to understand the content instead of mindlessly regurgitating it.

The more one outsources their thinking, the less practice they develop and maintain when it comes to thinking.

Universities have come up with novel solutions to get around simply handing in work authored by chatgpt.

That sounds like it might be effective at reducing reliance on LLMs in that context, but there's a lot of other contexts where people might need to rely on their ability to understand things, as well as their ability to tie together what they understand and draw conclusions based on those understandings.

People regurgitating things they've heard/read but don't understand isn't a problem that's only relevant in the context of universities after all.

Yeah I agree, students largely rely on rote learning and regurgitating information like ticking off a laundry list.

Not just students, because those students eventually leave school. Those people are the same way in their careers and private lives too.

That's what I meant by saying it isn't only relevant in the context of universities.

You've been involved in interactions where those sorts of people have been participants.

Sure. many professionals could turn to Grok, Chatgpt or any other LLM to do anything from final touches all the way to AI created reports. Of course the more they rely on the latter, the more management will wonder what value they add to the company or organisation.

Meanwhile average Joes can have all their questions answered in detail by a simple prompt. You might start seeing a growing number of unemployed white collar workers, but heck! you could also start seeing supersmart/knowledgeable butchers, bakers and candlestick makers.

Meanwhile China is using it extremely well professionally .-.