Learn the Systems, Teach the Machine, Stay the Author. ✍️

🔥

I have seen a few great posts now that all seem to be pointing to the same thing, and I have written a post a bit back that did not make it to the stage. But I think it needs to be at the forefront of how we are all working with AI. The Models you are using do not matter here; your critical thinking skills do!

I'm scared of something quieter; something being handed over freely that is one thing we can't afford to give away.

Our thinking.

There is data on what those costs. 👇

-------------------Learn the Systems. 🧠

You cannot direct what you don't understand. You cannot catch a mistake you can't recognize. You cannot verify an answer when you have no idea what right looks like.

So, learn the system first. The workflow. The architecture. The why under the what. Not so you can do every part by hand forever, but so you stay in command of the parts you hand off. Your system, your workflows, your responsibility.

When you understand the system, AI is a power tool. When you don't, it's a slot machine you keep pulling and hoping to win from. 🎰 I worked this way in the past, and until you step back from it, it's hard to see the flaws.

-------------------Teach the Machine. 🦾

Here's the reframe that changes everything 👇

You are not the one being assisted. You are the one doing the teaching.

You set the context. You write the spec. You define what good looks like. You correct it when it drifts. The AI is fast and tireless, and it has no idea what you want until you tell it, clearly, again and again.

(I designed a coach to help with this; I made it free because I believe it should be!) BuiltByBas/pre-flight-coach: Pre-Flight is an AI coaching and development workflow for new developers. The coach has been designed to sharpen design and structured thinking principles, it coaches you to clarity. Just download the pre-flight-coach folder and point Claude at it, the workflow will do the rest.

AI enhances. It does not replace. The moment you flip that and let it do the deciding while you do the watching, you stop being the teacher and become the student of a teacher that cannot think.

AI does not think, it does not understand, it does not feel, and it does not care if it gets it right or wrong. AI simply processes, it computes and it returns results based on YOUR instruction.

-------------------Stay the Author. ✍️

This is the one that matters most, and it's the one the research is lighting up. 👇

When people hand the thinking over, it shows. And now we can measure it.

A MIT Media Lab team (Kosmyna et al., 2025) wired people to EEG while they wrote essays.

Three groups: ChatGPT, search engine, and brain-only.

The ChatGPT group showed the weakest brain connectivity of the three. They reported the lowest sense of owning their own work. And many struggled to quote back the essay they had just written. The researchers called it cognitive debt.

And this is not an instant effect, it happens overtime, and the bill comes due later.

A Microsoft Research and Carnegie Mellon study (Lee et al., 2025) surveyed 319 knowledge workers across 936 real tasks.

The pattern was sharp 👇

1️⃣ The more someone trusted the AI, the less critical thinking they did.

2️⃣ The more they trusted their own skill, the more critical thinking they did.

AI didn't remove the thinking. It moved it, toward verifying, integrating, and owning the output. Toward staying the author.

A third study (Gerlich, 2025), 666 people, found the same shape. Heavy AI users scored lower on critical thinking, and cognitive offloading was the thing in the middle driving it. But here's the part I keep coming back to: the people with strong habits of reflection held the line. The skill protects you. The habit protects you.

-------------------One honest note, because overselling this would be its own kind of lie. 👇

None of these say AI makes you stupid. The loudest one, the MIT study, is still a preprint, run on a small group, and it already has a published critique asking for caution. The authors themselves asked people to stop saying "brain rot." The other two lean on self-report and correlation, not proof.

So I'm not telling you AI is the problem. I'm telling you the opposite. The cost doesn't come from using AI. It comes from checking out while you use it.

Every one of these studies points to the same exit:

Stay engaged, Stay skilled, Stay the author, and the cost drops.

-------------------So here it is. 👇

Learn the systems, so you can lead. Teach the machine, so it serves your intent. Stay the author, so the thinking stays yours.

Use AI every day. Be loud about it. I am. Just keep the pen in your hand. ✍️

We learn together, we grow together, we win together! 🤓💪🏆 Bas

-------------------Sources because they matter also! 😅

Kosmyna, N., et al. (2025). Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task. MIT Media Lab. Preprint, arXiv:2506.08872. https://arxiv.org/abs/2506.08872

Lee, H.-P., et al. (2025). The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers. Microsoft Research and Carnegie Mellon University. CHI '25. https://www.microsoft.com/en-us/research/publication/the-impact-of-generative-ai-on-critical-thinking-self-reported-reductions-in-cognitive-effort-and-confidence-effects-from-a-survey-of-knowledge-workers/

Gerlich, M. (2025). AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking. Societies, 15(1), 6. https://www.mdpi.com/2075-4698/15/1/6

5 comments