AI Empowers the Thinkers

The faculty training session started like most of them do. Seventeen people around a conference table, half of them already skeptical before I said a word.

I didn’t open with AI. I asked them about thinking. Specifically, I asked what was the hardest thing their students consistently got wrong — not because they lacked information, but because they hadn’t learned to think through a problem.

The room went quiet for about four seconds. Then it went loud for the next forty minutes. Every person at that table had a story, and the stories all had the same shape: students who could follow instructions perfectly but couldn’t navigate a moment when the instructions ran out.

We hadn’t touched AI yet. That was intentional.

Here’s the misconception I keep running into, in faculty meetings, in student surveys, in every panel discussion about the future of education. Everyone assumes the conversation about AI is a conversation about AI. It isn’t. It’s a conversation about thinking. AI just made the stakes visible in a way they weren’t before.

The real story is this: AI rewards structured, critical thinkers. It exposes everyone else.

I’ve been saying some version of this for a while, and I kept waiting for the research to contradict me. It hasn’t. A 2025 cognitive science study found that students who regularly used AI for recall and reasoning tasks showed measurable declines in those same skills when tested without it. That’s not a study about AI. That’s a study about what happens when you consistently outsource the hard parts of thinking.

The students who were already offloading their thinking before AI existed are the ones in trouble now. Not because AI made them worse. Because AI is finally making the problem visible.

Let’s talk about this for a moment.

We’ve been through this before. Every generation has had its version of this panic. Cliff Notes showed up in 1958 and English teachers were certain students would never read another book. Calculators arrived in classrooms in the 1970s and mathematicians worried we’d produce a generation that couldn’t add. The internet came along and everyone was sure research skills would evaporate overnight.

What actually happened? In each case, the students who were genuinely engaging with the material found that the new tool made them sharper. The students who were already looking for the shortcut found a faster route to the same empty destination. The tool didn’t change the destination. It just shortened the trip.

AI is the same pattern. A different order of magnitude, but the same pattern.

Here’s what I’ve noticed about students who use AI well. They do five things consistently that the shortcut-takers don’t.

1. They bring the question, not just the assignment. A thinker asks AI to help explore a problem. A shortcut-taker asks AI to solve it. That difference in framing determines almost everything about the quality of what comes back.

2. They verify everything. Not because they distrust AI specifically, but because verification is part of how they approach every source. Thinkers were checking sources before AI existed. They’re still checking sources now. A preprint study from Princeton and the University of Washington found significant commercial bias in multiple leading AI systems — and that most users never thought to check. That’s not an argument against using AI. That’s an argument for bringing your critical thinking to every session.

3. They know what good looks like before they ask. You cannot evaluate an AI-generated design, argument, or analysis if you don’t have a working model of quality already in your head. Thinkers build that model through study and practice. They use AI to extend and pressure-test it.

4. They treat AI as a collaborator, not an oracle. A collaborator can be wrong. You push back, you refine, you test the ideas. An oracle is just obeyed. The students who obey AI are producing work that looks like something until you examine it. The students who collaborate with AI are producing work that holds up.

5. They own the thinking, not just the output. AI doesn’t have professional judgment. It doesn’t carry your experience, your knowledge of what failed last semester, your understanding of what the client actually needs underneath what they said. Those things live in you. The thinkers bring those to the session. Everyone else just hopes the output looks right.

Here’s something students eventually discover on their own, usually after a painful moment. The more you actually know about a domain, the more useful AI becomes. That sounds obvious until you sit with it. AI doesn’t replace expertise. It multiplies expertise. If you bring nothing to the conversation, you get output that looks like something. If you bring genuine knowledge and structured thinking, you get a force multiplier.

That’s not a metaphor. It’s the mechanics of how these tools work. The quality of your input shapes the quality of your output. Without exception.

This is why the Good-Cheap-Fast triangle still applies in the age of AI. You remember the rule: Good, Cheap, Fast. Pick two. AI broke the cost curve. AI absolutely changed the speed equation. But it didn’t change what “good” requires. Good still requires judgment. Good still requires criteria. Good still requires someone who knows the difference between a result that works and a result that just looks like it works.

AI can give you fast. AI can give you cheap. The good is still on you.

I told that faculty group the same thing I’m telling you now. The question isn’t whether your students are using AI. They are. The question is whether they’re using it as a thinking tool or as a thinking replacement. Those two things are not the same. They don’t produce the same work. They don’t prepare students for the same futures.

You can build a practice around that distinction. Design your projects to require demonstrated judgment, not just completed output. Teach verification as a skill, not an afterthought. Model what structured critical thinking looks like when applied to an AI conversation — and what genuinely useful output looks like when that thinking is present.

The students who are already thinking carefully are going to figure this out with or without your help. They usually do. But the students in the middle — the ones who could go either direction — those students need someone to name it for them. To show them the difference between using a tool and leaning on it.

That’s the job. It was always the job. AI just made it more urgent and more visible.

The thinkers in your classroom are going to be unstoppable. Build the environment that produces more of them.