Own Your Thinking Before AI Does
If you don’t own the model in your head, the model will own you
A colleague pinged me on a Tuesday afternoon.
She was deciding whether to restructure her team — a real decision, with real stakes, two weeks of back-and-forth and no clear answer. She trusted my read on these things. So she asked.
I opened AI. Typed in the context. Ninety seconds later I had something clean: four considerations, a recommended path, even the framing she should use with her manager. It was good. Structured. It even used a couple of phrases I probably would have chosen myself.
I read it. Sent it.
Three days later she came back. “I tried your framing. My manager pushed back and I didn’t know what to say.”
I didn’t know what to say either. Because I hadn’t actually thought it through. I’d outsourced the thinking, packaged the output, and called it advice. The answer was clean. The conviction behind it was zero.
She didn’t get bad advice. She got borrowed advice. And there’s no way to defend borrowed thinking when someone starts pulling on it.
The Shift You Don’t Notice
When AI gives you bad output, you know immediately. You push back. You catch it. You stay in charge.
This is about what happens when the output is good.
When it’s coherent and settled and sounds like someone already figured it out — your brain does something convenient. It stops working. Not because you’re lazy. Because the output fills the gap before your own thinking has a chance to show up.
The drift is slow. It starts with “help me write this email.” Then “help me think through this.” Somewhere along the way it becomes “just tell me what I should do here.” You don’t notice the slide because the outputs keep being useful.
And then someone pushes back. And you’re empty.
You’ve had a version of this. Maybe not with a colleague. Maybe in a meeting, answering a question with a response that felt borrowed. Maybe publishing something that sounded like you but came from somewhere else.
The tool didn’t fail you. You stepped back too far. And the longer you keep stepping back, the harder it gets to find where you left off.
Three Stages. One Rule.
I spent the better part of last year working out how to stay the author of my own ideas without dropping tools that genuinely make me sharper. What I landed on is three stages.
Play. Pattern. Production.
Most people skip the first two. That’s the problem.
Play
Play is where you use AI to widen the field.
The default move is to ask for the answer. What should I do? What’s the right call? How should I think about this? That’s handing over the wheel before you’ve decided where you’re going.
Play looks different. You ask for terrain instead. “Give me five different ways to think about this.” “What’s the strongest argument on each side?” “Where do smart people usually get this wrong?” You’re not looking for a conclusion — you’re building more ground for your own thinking to push against.
One rule: no conclusions. Not yet. You’re generating raw material — and what you build from it is entirely your job.
Pattern
After a Play session, close the tab.
Not forever. Just long enough to notice what’s still in your head. Which ideas keep coming back. Which argument bothered you. Which answer felt close but slightly wrong — and you can’t explain why yet.
That dissonance is your thinking trying to show up.
Good AI output sounds settled. Sounds like someone already figured this out. And if you move straight from Play to Production without stopping, you end up saying what the AI said, with your name on it.
Pattern is where you find your actual position. Three sentences. Messy is fine. You’re not trying to be articulate — you’re trying to be honest. What do I keep returning to? What am I resisting? Where does my gut disagree with the logic?
Write it down. Then open the tab.
Production
Now you bring the argument. Now AI comes back in.
This is where most people start. That’s the mistake.
Production only works when you already know what you’re trying to say. You arrive with a real position. Rough around the edges is fine. Then you use AI to organize it, sharpen it, fill the gaps.
The question shifts completely: “Help me say this more clearly.”
One produces owned thinking. The other produces borrowed thinking with your name on it. The difference shows up the moment someone pushes back — because you either built the argument or you accepted one.
Before You Open the Next Tab
Play to think wider.
Pattern to find yourself in the material.
Production to say what you actually think.
Before your next AI session, try this first: write three sentences of your raw take. Don’t make them good. Don’t worry about being right. Just make them yours. Then open the tab.
You’ll notice it immediately. When you already have a position, the output hits differently. You’re not looking for an answer anymore. You’re stress-testing one.
Before you open the next tab, make sure you have something worth testing — not something worth accepting.
What’s something you actually think — that you haven’t let AI clean up yet?
Reply and tell me. I’m genuinely curious.
P.S. This is what I’ve been working through — how thinking without architecture defaults to borrowed conclusions. Not just with AI. Everywhere decisions get made. More on that soon.