Three Wrong Conversations About AI in Education (and the One We Should Be Having)

I've spent the last year talking to educators, parents, and tech people about AI in schools. Three conversations keep coming up. All three are wrong.

1. The Cheating Panic

"Students are using ChatGPT to write their essays. We need to shut it down."

This conversation assumes the only valid form of assessment is: student produces a deliverable alone, teacher grades it. If that's your entire model, then yes, AI breaks it. But that model was already broken. It just took AI to make it obvious.

The panic misses something important. AI enables entirely new forms of assessment that are actually harder to game. A student explaining a concept to an AI in a Socratic dialogue can't fake understanding. A student teaching the AI — where the AI plays dumb and asks follow-up questions — reveals gaps in knowledge that a polished essay never would. An AI that pushes back on your reasoning forces you to think, not perform.

These approaches require imagination and redesign. Banning AI is the path of least imagination.

2. Speedrunning the Checklist

"AI can help students master the standards faster and more efficiently."

This sounds reasonable until you think about what it actually means: using the most powerful technology in a generation to speedrun a fixed checklist. Cover more Common Core standards per hour. Raise test scores. Move through the curriculum faster.

This is the turbocharger approach. Same engine, same direction, just faster. Efficient coverage, adaptive practice, rapid skill acquisition. And it works for a lot of kids — especially kids who are motivated by points, streaks, and leaderboards.

But this approach leans heavily on extrinsic motivation. For curious, intense kids who want to go deep on things that interest them — kids who ask "why?" more than "what's on the test?" — speedrunning the checklist faster doesn't solve the problem. The problem was the checklist.

The better question: is the existing curriculum even asking the right things?

3. The Vocational AI Skills Track

"We need to teach kids AI skills so they can get jobs in the AI economy."

I founded an AI startup. I've been in tech for 20 years. And I can tell you: nobody in industry knows which specific AI skills will matter in five years. The tools change every few months. The workflows change every few months. The jobs themselves are being redefined in real time.

Training kids on specific AI tools is like training them on specific software packages in the '90s. Remember Lotus 1-2-3? HyperCard? The schools that taught "computer skills" by drilling kids on particular applications were preparing them for a world that stopped existing before they graduated.

Even coding — the skill most associated with tech fluency — is being fundamentally transformed. Much of what engineers used to do is now handled by AI. The people who thrive aren't the ones who memorized syntax. They're the ones who can think clearly about what needs to be built, evaluate whether it's working, and communicate precisely about what's wrong.

The Conversation We Should Be Having

All three wrong conversations share the same mistake: they start from the existing system and ask how AI fits into it. How do we protect traditional assessment? How do we accelerate traditional curriculum? How do we add AI to the traditional career pipeline?

The real question goes the other direction. Start from what we know about human development and ask: what capacities will matter regardless of how the technology changes?

Three stand out.

Clear, precise communication. The ability to articulate what you mean, recognize when something is off, and refine until it's right. This was valuable before AI. It becomes essential when your primary tool responds to language.

Systems thinking. The ability to reason about complexity, see how parts connect, understand second-order effects. As AI handles more routine cognitive tasks, the ability to think about whole systems — not just components — becomes the differentiator.

Empathy and democratic dispositions. Care for others. The ability to work with people who think differently. A sense of responsibility to community. These are the capacities that AI will never replace, and the ones that matter most in a world where technical skills have a shorter and shorter shelf life.

Once you're oriented around these capacities, AI stops being a threat to manage or a tool to bolt on. It becomes something more interesting: a way to cultivate these habits of mind.

An AI Socratic partner doesn't just deliver content faster. It creates a space for deeper thinking, reflection, and dialogue. A "teach the AI" exercise doesn't just assess knowledge. It develops communication skills in real time. An AI that challenges your reasoning doesn't just check answers. It builds the capacity to engage with complexity.

This is what we're building at Tradewinds. Not AI that makes school more efficient. AI designed to develop the human capacities that will matter no matter what comes next.

Why Openness Matters

There's one more thing worth saying. When AI mediates education — when it shapes which questions get asked, how concepts get scaffolded, what feedback a student receives — the pedagogy lives in the prompts. The editorial point of view is baked into the system. Every design choice about how the AI responds to a student is a pedagogical choice.

That means AI in education needs to be inspectable. Parents should be able to see what's guiding their child's learning. Teachers should be able to shape it. Researchers should be able to evaluate it.

The alternative is a black box controlled by a vertically integrated company that sells the curriculum, the platform, the assessment, and the AI, all in one bundle. You can't see the prompts. You can't change the pedagogical approach. You can't even tell what values are embedded in the system your child spends hours with every day.

We build in the open. Our AI tutor's prompts, pedagogical framework, and decision logic are visible to the teachers who use them and the families who trust us with their kids. When something isn't working, we can see why and fix it together. That transparency is a prerequisite for trust.