Key takeaways
- The MIT Media Lab preprint 'Your Brain on ChatGPT' (54 people, EEG caps, essay writing) found chatbot users had the weakest brain connectivity, the lowest sense of ownership, and couldn't quote their own essays.
- It's an early preprint on a small, narrow task. It does not show AI 'rots your brain,' and the researchers themselves cautioned against that reading.
- The real lesson is about a habit: hand the hardest thinking straight to a machine and you end up unable to defend work that was never really yours.
- For a coach, consultant, or advisor, your judgment is the product. Clients pay for your read on their situation, not for a document they could generate themselves.
- Split every task into thinking (the decision, the argument, what to cut) and packaging (drafting, reformatting, proofing). Keep the first, automate the second.
- Do your thinking first in messy form, then use AI to package it, with an explicit rule that it can't add any conclusion you didn't write.
- When you don't have a view yet, have AI interview you instead. It may only ask questions, never answer, so the argument that comes out is still yours.
- The one-question test: is this a decision or a delivery? Decisions stay with you. Deliveries go to the machine. If you can't quote your own point afterward, redo the thinking yourself.
The Solo Creators AI Studio
Get this skill, and every one I build.
The whole vaultA new drop monthlyBuild live with meYours to keep, forever
See inside the Studio
30-day money-back guarantee. Cancel anytime.
Questions people ask
Does using AI actually make you worse at thinking?
The MIT Media Lab study found that people who leaned on a chatbot to write essays showed weaker brain connectivity and couldn't quote their own work, but it's a small early preprint on one narrow task. It doesn't prove AI permanently damages your thinking. What it points to is a habit: if you hand over the hardest part of the work and copy back the output, you never build ownership of it. Use AI for the labor and keep the judgment yourself, and that risk mostly disappears.
What did the MIT 'Your Brain on ChatGPT' study really find?
Researchers put EEG caps on 54 participants and had them write essays with a chatbot, with a search engine, or with no tools. The chatbot group showed the weakest brain connectivity, reported the least ownership of their essays, and struggled to quote lines they'd just written. The no-tool group showed the strongest, most connected activity. It's a preprint, not a peer-reviewed final paper, and the authors were careful to say it doesn't support alarmist 'brain rot' conclusions. You can read it on the MIT Media Lab site.
So should I stop using AI for my work?
No, and I don't. I build my whole business with it. Stay clear about what you hand over. Automate the packaging: drafting from your notes, reformatting one idea into an email and a post, proofreading, transcribing. Keep the thinking: the diagnosis, your argument, what to recommend, what to cut. Hand over the labor, keep the decisions.
What's the difference between 'thinking' and 'packaging'?
Thinking is any task where the answer depends on your read: what the client's real problem is, what you recommend, what to leave out, whether the draft is even right. Packaging is turning a decision you've already made into a usable shape: a clean draft, a shorter version, ten subject lines, fixed typos. Ask 'is this a decision or a delivery?' Decisions stay with you. Deliveries go to the machine.
How do I use AI to write something without letting it think for me?
Do your thinking first in whatever messy form it comes out, a voice memo or bullets, then hand that raw material to AI with strict rules: it can only shape what you wrote, it can't add any conclusion you didn't include, and if it spots a gap it has to hand you a question instead of inventing an answer. There's a copy-paste prompt in this article that does exactly that. You stay the author and still save the hour.
What if I don't have a point of view on the topic yet?
Then don't ask AI for one, because that's the moment you hand over the part that's actually yours. Flip it around and get interviewed instead. There's a second prompt in this article that tells AI it may only ask you questions, one at a time, and may never offer an opinion or a recommendation. By question six or seven you'll usually hear yourself say what you actually believe, and then it plays your answers back as your own argument. Now you have a view you can defend, and you can run it through the packaging prompt.
Why does ownership of my own work matter so much?
Because for a coach, consultant, or advisor, the product is your ability to defend a recommendation under pressure. A client asks 'why this and not that,' and either you did the thinking yourself and can answer, or you copied it from a chatbot and you're bluffing. The study's finding that chatbot users couldn't quote their own essays is the exact failure mode. You can't fake ownership in a live room.
Written by Ronnie Nijmeh
Ronnie spent 18 years building a SaaS with a team of 20 that served over 650,000 customers, generated over $14M in sales, and sent over 550M emails. Now he's solo, solving real business bottlenecks and turning them into working AI skills, workflows, and automations. He teaches all of it, with direct access to him, inside the Solo Creators AI Studio Skool community. See what he's built →
“I completely rebuilt my site using the Website Launch Builder. If you've ever built a website, you'll really appreciate this skill. I'm blessed to be a member.”
Michael HattawayFounder, Iron Strengthens IronKeep reading
CLIENT-READY AI
Send Work That Lands: Why AI Output Becomes Someone Else's Homework
Polished AI output that reads done and isn't quietly hands your client the real work. Learn the tell, and the fix, so your deliverables land the first time.
AUTOMATION STRATEGY
Automate the Right Half: What Klarna and Duolingo Learned the Hard Way
A clear decision table for solo creators: automate the packaging around your work, never the judgment that is the work.
PRODUCTIVITY
Find Out If AI Really Saves You Time: A One-Week Measurement
A skeptic-proof, one-week protocol for measuring whether AI genuinely saves you time, built around the one study that caught people being confidently wrong about their own speed.