are you still there?

How to protect your brain from AI (before it turns into jello).

Last week, MIT dropped a study that basically confirms what we’ve all been feeling but trying not to say out loud: using ChatGPT too much is making us stupid.

Like, for real stupid. Not in a haha I forgot my password again way — in a low neural engagement, poor memory retention, zero creativity way.

Here’s what the MIT peeps actually did:

They took 54 participants between 18 and 39 years old and split them into three groups. One wrote essays solo (no tools), one used Google Search, and one used ChatGPT. Each group completed multiple SAT-style writing tasks — 20 minutes each, on prompts like the ethics of philanthropy or the paradox of choice. Meanwhile, their brains were being monitored with EEGs across 32 regions.

The ChatGPT group had the lowest neural engagement across the board. Executive control? Low. Attentional focus? Low. Linguistic diversity? Minimal. Creativity? Flatlined. By the third essay, most were offloading the task entirely — feeding prompts to the AI and pasting the output with barely a tweak.

Worse, when asked to re-write one of their own essays (without AI), they couldn’t remember what they had written. Their brainwaves showed weaker alpha and theta activity — the kind that signals deep processing and long-term memory formation. In other words: they didn’t just underperform. They didn’t integrate any of it. The task got done, sure. But it left no trace.

but isn’t it… efficient?

Sure. If the goal is to complete the task, get the dopamine hit, and move on. But if the goal is to think? To process, remember, connect dots, develop taste, get better? It’s giving brain fog with wifi.

One of the most unsettling things the researchers found was that when ChatGPT users were later asked to re-write their own essays — without help — they remembered almost nothing. The content never made it into long-term memory. No alpha or theta waves. No creativity. Just blank space.

It's literally like asking someone what they ate for dinner and them going: "I... outsourced it."

ok so now what.

This isn’t a TED Talk. There’s no “5 easy ways to fix it.” But if you’re like me — still using AI daily, but not trying to become its content funnel — here’s what I’m doing to keep my brain from turning to soup:

→ write cold
I don’t let AI near my first drafts. That’s sacred ground. If I let the bot jump in too early, I end up reverse-engineering someone else’s voice instead of following my own. It’s like trying to find your rhythm while someone hums in your ear — helpful, maybe, but now it’s their melody, not yours. The first draft is supposed to be chaotic, stupid, too long. That’s how you know it’s real. That’s how you know it’s you.

→ boredom is back
I’m re-learning how to be bored. Not “scroll until your eyes hurt” bored. Actual, look-out-the-window-and-hate-it boredom. The kind that makes you notice things. That makes ideas show up uninvited. It’s awful. But that’s kind of the assignment.

→ set traps
Before I open any AI tool, I give myself a prompt: what am I actually trying to do here? If I don’t have an answer, I don’t open the tab. No vibe-based prompting. No browsing for epiphanies. My brain can wander — theirs can’t. That’s the line.

→ make memory a muscle
I stopped pretending everything’s “saved.” If I read something and can’t recall it tomorrow, it’s gone. That’s on me. So I force recall. I narrate out loud. I write from memory. I let ideas decay, just to see what survives. Not everything needs to be retrievable — but something should be retainable.

→ ask better questions
AI has answers. Too many. Most of them fine. Which makes them worthless. The only leverage left is in the question: why are you asking? What are you really trying to find out? What would be risky to know? Asking better isn’t just a skill now — it’s the only way to cut through the flood of okay.

last thing.

This isn’t an anti-tech piece. I use Chat GPT, Perplexity, etc. every day. I love what it enables.
But I also love my mind.
And I’d like to keep it.

The risk isn’t that AI becomes sentient.
It’s that we stop being fully conscious — and don’t notice.

So yeah. I’m scared.
But I’m still thinking.
Hope you are too.