What Brainrot Actually Is (And What It Isn’t Yet Proven To Be)

I have a good memory. Always have. Names, numbers, the specific thing a client said in a meeting four months ago – I could usually pull it back without trying. So when I started forgetting small things, I noticed, but not in any dramatic way. I’d lose my train of thought mid-sentence. I’d forget something I’d read the week before, or something someone had told me directly, looking at me, expecting me to remember it.

None of it was alarming on its own. That was the problem. It was small enough to explain away every single time.

I only saw the shape of it once it was gone. Three weeks ago I deleted every social app off my phone, and somewhere in that stretch, the fog I hadn’t fully clocked started lifting. Which means I’d been carrying it for a while without knowing what it was, or how much of it was actually a choice.

That word – brainrot – has been sitting in the middle of this. Oxford made it their word of the year. Everyone throws it around now, usually as a joke, sometimes as a genuine worry. I wanted to know what it actually is before I used it about myself. What I found is messier, and more honest, than the version that’s currently going viral.

Oxford put a word on it. That doesn’t mean the science agrees.

Brain rot was Oxford’s Word of the Year for 2024. Their definition: “the supposed deterioration of a person’s mental or intellectual state, especially viewed as the result of overconsumption of material – now particularly online content – considered to be trivial or unchallenging.” Usage of the term jumped 230% in a single year. Over 37,000 people voted for it in Oxford’s public poll.

That’s a real cultural moment. It is not, on its own, evidence of anything happening inside anyone’s skull.

Here’s the part that gets skipped in most of the coverage: an Oxford University professor has gone on record saying there’s no evidence brain rot is an actual neurological phenomenon. Not “the jury’s out” – no evidence, stated plainly, by someone at the same institution that just crowned the term word of the year. A media psychologist quoted elsewhere put it just as bluntly: the evidence that screen use harms intelligence is still weak, and there are very few studies that scientifically support the idea in the form people are using it.

So you’ve got a word winning popularity contests at the same time actual researchers are telling you to slow down. That gap is worth sitting in for a second, because it tells you something the definition itself doesn’t: this is a term that arrived faster than the proof did.

What the evidence actually shows, and where it runs out

I’m not going to tell you nothing’s happening. Something is. I just want to be honest about what the something actually is, because most of what’s written about this either overclaims or dismisses, and neither one is doing you any favours.

The strongest single piece of evidence right now is a 2025 meta-analysis published in Psychological Bulletin, pulling together 71 studies and 98,299 people, specifically on short-form video – TikTok, Reels, Shorts. Heavier use tracked consistently with weaker attention, weaker inhibitory control, weaker working memory, plus more depression, anxiety, and loneliness. That’s a serious body of evidence.

It’s also correlational, and the researchers say so themselves – people who already struggle with focus may simply be more drawn to short-form video in the first place. The study shows the two things travel together. It doesn’t prove which one is pulling the other.

Then there’s the study that speaks directly to the AI half of this, and it’s the one I find most useful, with the biggest asterisk attached. MIT Media Lab ran a study – “Your Brain on ChatGPT” – watching what happens to people’s brains while they write essays with an AI assistant versus a search engine versus nothing at all.

Using EEG to track brain connectivity across four months and multiple sessions, the group using the AI assistant showed the weakest neural connectivity of the three, and performed worse – not just while using the tool, but afterward too, when the AI was taken away. That’s the closest thing I’ve seen to direct evidence that leaning on AI to think for you leaves something behind once you stop.

Here’s the asterisk: it’s a preprint, not yet peer-reviewed, and the sample shrank from 54 people down to 18 by the final session. That’s not nothing. It’s also not proof. Treat it as the most promising early signal we have, not a verdict.

That’s the honest state of the evidence. Real associations, real early signals, and a lot of space between “something is happening” and “we know exactly why.”

Why “doomscrolling causes brainrot” is the lazy version

This is the bit that actually irritates me, and it’s the reason I wanted to write this instead of just sharing a study.

The popular version of brainrot has one villain and one mechanism: you scroll, your brain rots. It’s tidy. It’s also, I think, wrong in the way that half-true things are wrong – not because scrolling is harmless, but because pinning it all on one behaviour lets everyone skip the harder, more tangled explanation.

Here’s what I think is actually going on, and I’ll say upfront – this is my read of the evidence, not a settled finding, because nobody has actually studied all these pieces together yet.

Short-form video is built around instant gratification. Each clip is a small hit before your brain’s own reward chemistry has time to settle, which is why one video turns into forty without you deciding that’s what you wanted. That part’s well understood. But it’s not just the scrolling mechanism doing the damage – it’s what’s actually in the feed, and how constantly you’re measuring yourself against it. Comparison is baked into the format. You’re not just watching content, you’re watching other people’s highlight reels on a loop, and that’s a different kind of tax than passive consumption alone.

Then there’s AI, and this is the piece most brainrot takes leave out entirely, because the word got popular before AI assistants were something most people used daily. I think heavy AI use trains a kind of dependency – not on a specific app, but on outsourcing the effortful part of thinking. A task that used to take a few hours now takes minutes, and the temptation is to treat that as a pure win.

But if the tool is doing the reasoning, you’re not doing it, and the research on cognitive offloading backs this up directly: studies have found a real negative relationship between how often people lean on AI tools and how they score on critical thinking, with that “offloading” effect doing a lot of the explaining. The MIT study above is the sharpest single data point for this, but it’s part of a wider pattern – hand the thinking over enough times, and the muscle for doing it yourself doesn’t get used.

Put those two together – a format engineered for constant reward, and a habit of letting a tool carry your thinking – and you get compounding, not addition. Neither one alone explains what I noticed in myself. Both together, over months, explains it a lot better.

Nobody’s proven that specific combination yet. I don’t think that makes it wrong. I think it means the researchers haven’t caught up to how people actually live now, using both things at once, every day.

We’ve panicked about this before

If “brainrot” feels like it arrived out of nowhere, it’s worth knowing it didn’t, and neither did the panic underneath it.

Thoreau used the phrase “brain rot” in Walden, in 1854, complaining that people preferred simple ideas to complex ones. Wordsworth made basically the same complaint about novels in 1800 – too much frantic, trivial reading, rotting the mind. Jump forward to the 1940s and a Journal of Pediatrics study described three-quarters of the children studied as “addicted” to radio and movies. Frederic Wertham spent the 1950s building a moral panic around comic books, convinced they were driving juvenile delinquency – it worked well enough to reshape the entire comics industry. Television got the same treatment a generation later.

Every new format that captures attention gets accused of rotting the minds using it. Sometimes the concern turns out to be onto something. Often it’s mostly a generation being uncomfortable with how the next one spends its time, dressed up as science.

I don’t think that history proves today’s version is nothing. I think it means “everyone’s saying it, so it must be true” has been wrong often enough that it shouldn’t be the bar. The APA meta-analysis and the MIT study are real evidence in a way Wertham’s comic book theories never were – but the instinct to reach for a single dramatic explanation, and to be very confident about it very fast, is exactly the same instinct that’s been wrong every previous time it showed up.

What I actually did, three weeks in

So here’s where I stop reporting on other people’s research and tell you what I actually tried, because I think that’s more useful than another paragraph of hedged evidence.

I was reading Johann Hari’s Stolen Focus, and somewhere in it the idea took hold: stop guessing, and just test it. So I deleted every social app off my phone. Not the accounts – the apps. Three weeks ago now, and I haven’t put any of them back.

I didn’t just remove something, though. I replaced it. I started keeping an ideas book – a physical place to write down thoughts, half-formed ones, questions I wanted to come back to, instead of reaching for my phone the second my brain had a gap to fill. That part matters more than the deletion, I think. Removing the scroll just opens up dead time. What you put into that dead time is what actually decides whether anything changes.

Three weeks in, I’m thinking more clearly. The ideas are coming faster, and there are more of them – not because I’m trying harder, but because there’s actual space for them to show up in. And something softer than that, harder to put a number on: I feel more settled. More content. I’m not going to dress that up as a measured outcome, because it isn’t one. It’s how I feel, reported honestly, three weeks into something I’m still in the middle of.

I’m not telling you this because I think I’ve cracked it. I’m telling you because it’s the only part of this whole conversation I can actually vouch for without hedging.

Find your own test

I don’t think the answer is a universal one. “Delete everything, replace it with a notebook” worked for me over three weeks – it’s not a formula, and I’d be doing exactly the thing that annoys me about the rest of this conversation if I sold it to you as one.

What I’d actually say is this: the researchers don’t have this fully figured out yet, and neither does the internet, whatever confidence the word “brainrot” seems to carry. What you do have is the ability to test it on yourself, the way I did – notice what’s actually filling your time, take one thing out, put something deliberate in its place, and pay attention to what changes. Not forever. Just long enough to see.

Somewhere between doomscrolling and never touching a screen again is an actual life, being lived. Finding your version of that balance is the whole point – not proving a theory right, just getting more of your own attention back.

Where you start is up to you. Mine started with four deleted apps and a notebook I almost didn’t buy.

FAQ

Is brainrot an actual medical diagnosis? No. It’s not recognised as a clinical condition – it’s a cultural term, not a diagnosis a doctor would give you. Real cognitive effects have been associated with heavy screen and AI use in research, but “brainrot” itself describes a pattern people recognise in themselves, not a medical category.

What’s the difference between brainrot and doomscrolling? Doomscrolling is a specific behaviour – compulsively scrolling through negative or anxiety-inducing content. Brainrot is broader and describes the supposed mental fog or decline that behaviour is blamed for causing. Doomscrolling is one possible contributor among several, not a synonym for the outcome.

Can brainrot be reversed? There’s no formal study proving reversal, but plenty of people, myself included, report noticeable improvement after cutting back on the behaviours associated with it. The honest answer is that nobody’s run a controlled study proving a fix – what exists is early evidence and a lot of self-reported experience.

Is AI worse for your brain than social media, or are they the same problem? They’re different mechanisms causing a similar outcome. Social media leans on fast, repeated reward loops. AI leans on cognitive offloading – letting a tool do the thinking you’d otherwise do yourself. Used heavily together, they likely compound each other, though no study has directly tested that combination yet.

What is cognitive offloading? It’s the practice of using an external tool to do mental work you’d otherwise do internally – looking up a fact instead of remembering it, or letting an AI assistant reason through a problem instead of working through it yourself. Some offloading is normal and useful. Research suggests heavy, habitual offloading is linked to weaker critical thinking over time.

Does brainrot only affect younger generations? No. The concern gets aimed at Gen Z and Gen Alpha most often because the term originated in those communities, but the underlying behaviours – heavy scrolling, heavy AI reliance – aren’t age-locked. Anyone with a phone and an AI assistant is exposed to the same mechanisms.

What should I replace scrolling time with if I want to test this myself? Something that asks something of your attention rather than absorbing it passively works best – a notebook for ideas, reading, a hobby you’d abandoned. The point isn’t the specific activity, it’s giving your brain something to do with the space you’ve freed up, rather than leaving it empty.