How do you know when you’re collaborating with AI, or outsourcing your judgment to it?

PART 1 — The day AI almost killed my TEDx talk and what saved it

I was already convinced that the most important thing when working with AI is not to forget your own mind into the tool. Then something happened that proved it. This is that story.

I had everything. Two articles, months of research, pages of thinking notes, a thesis I believed in deeply. When my talk got selected, the idea was formed. What remained was the hardest part: shaping all of that into something a room full of strangers could follow, feel, and remember in under twelve minutes.

TEDx has strict constraints. Short format. Strong structure. A thesis that matches the event’s theme. A timeline that gave me just over a month from selection to stage, including memorization, embodiment, rehearsal, and the work of actually living the thing out loud.

I had fears working against me from the start. I mean, who wouldn’t be a bit intimidated by standing on a TEDx stage? What I didn’t see was what they were doing to my thinking underneath.

I gave AI everything. Every note, every source, every half-formed idea. I thought it would speed things up.

What came back was technically structured and… said nothing. I went deeper into the loop. I kept generating, kept checking, kept asking for different framings. At some point I caught myself thinking: what am I actually doing here? Am I even capable of assembling this? Every output looked coherent in some way and I couldn’t tell what deserved my attention anymore. I started wondering if I was losing my capacity to struggle through something, and whether that mattered. And then the thought that stopped me cold: if AI writes this for me, am I even the expert I say I am?

That last question broke the loop. Because it forced me to notice what had been happening. The fears I hadn’t examined were running the process. AI was giving them a very efficient surface to work on. The bottleneck was selection, not production. And selection required a human act: trusting my own judgment enough to choose. Somewhere between the pressure, the constraints, and the tool’s speed, I had lost it.

That moment of noticing is called metacognition.

But the noticing didn’t happen at the desk. It happened later. In the silence I had made space for, one of those moments of doing nothing that I’ve learned, over years, to protect. Sitting. Staring at the wall. Letting my mind wander without agenda.

At some point, my mind started asking questions on its own. Why am I so stuck? Why does it feel like I’m not getting anywhere? Am I even gonna make it? And then, following the thread: what am I actually scared of?

I was scared of standing behind an idea. Scared of not being taken seriously. Scared of saying something wrong. And when I compared those fears to what I had been doing with AI — constantly regenerating, rechecking, asking for better formulations, not trusting myself — the loop made complete sense. I hadn’t been collaborating with AI. I had been outsourcing my fears to it.

That moment was only possible because I had done years of work learning how my mind behaves. My emotional patterns, my shortcuts, my thinking style, the specific ways I avoid discomfort. Without that self-knowledge, I would probably have kept generating. The loop would have looked like a workflow problem, something that sounds like: “AI is not smart enough” or “I don’t prompt it well”. It wasn’t.

I had thought about it already. The talk I was trying to write was even exactly about it. What we do shapes what we become, and what we do is shaped by how we think. If you’re not paying attention to your own mind, something else is doing it for you. I was living exactly what I was talking about — and it became easier to write, not from something I knew intellectually, but from lived experience. From that moment on, I understood something I now believe every person working with AI needs to reckon with: staying in charge means monitoring and regulating your own thinking while using the tools. Since then, I researched, observed, talked to people about their experience. Now I’m convinced. Next week, I’ll tell you why metacognition is the skill of the age of AI.

If this landed and you don’t want to wait until next week, or until you prove it on yourself the hard way, I built something for that. Own Your Thinking is a one-day intensive for teams who want to stay the author of their decisions as AI becomes part of the process. 

Know more about Own Your Thinking solution.