The Machine Has Always Been in the Room
The fear is real and actually well placed.
I talk with the teachers I work with about AI all the time, and my argument is the same to those that embrace it and those that see it as the enemy: AI is a tool that we have to learn to use to improve human thought and experience, the same way people sounded the alarm on calculators, on the internet, on search engines. I still hold that to be true. But AI is more than a tool at this point, and it will only grow further.
A better way to describe AI right now is that it is an infrastructure. It does way more than write essays and answer questions, and there is a real, growing subsection of the population that believes AI should be teaching our students instead of humans. But what would AI do better than teachers that would justify replacement? Teaching the content? We have already transitioned from textbooks to slideshows to websites to all kinds of different transport mechanisms for information. The tool has just changed. At the end of the day, to relinquish the educational experience of a child into the hands of an artificial intelligence model — where they communicate solely through a chat interface — would be harming them. Interaction with a human being is something that is learned through practice.
And that's not a philosophical point. It's a developmental one. School was never only about content delivery. It was always about socialization, conflict resolution, learning to sit in a room with people you didn't choose and figuring out how to exist together. A chat interface cannot replicate that. It doesn't even try to.
Teachers are already being shaped by AI. That's the honest truth. It is a huge effort to reach burnt out, underpaid, and overworked teachers with the theory of AI and have real conversations about how to shape it. Couple that with unscrupulous SaaS companies trying to make a quick buck by building pedagogically sub-optimal software, and you aren't bound to get very far. We need to have serious conversations and we need to have them fast. We need to be at the forefront of discussions around AI, because teachers aren't only educating their students — they are educating everyone that uses these tools. Who is going to teach people how to recognize hallucinations and AI-generated images? Teachers should. That's our job.
Right now, every app I have seen put in front of teachers and students has been a shell of what it could be. The trend is very quickly built applications that plug into AI APIs — essentially wrappers for generative AI models that you can access through a chat interface. You can accomplish the same goals and outcomes as a lot of these AI companies if you just teach teachers how to prompt the right way. Some teachers are catching on, but there is a severe disconnect between what AI can actually do and what educators are using it for. The game-based web apps are oddly the ones doing it well, because game designers have always had to solve the same problem good teachers solve — how do you keep someone engaged with something hard? How do you make failure feel like information instead of judgment? The difference is that a good game designer builds the feedback loop around the learning objective. A bad edtech company builds the loop around the session length. Same mechanic. Completely different ethics.
If I could sit every edtech CEO in a room and say one thing, it would be this: if you got into educational technology in an attempt at financial gain, you are missing the point of what education has been, is, and should be. There are ways to make money by releasing things for free. It is harder, sure, but there are ways. Education is not an industry to be entered into. It itself is an infrastructure. Just like AI is becoming an infrastructure. And when you treat infrastructure like an industry — when you privatize it, when you optimize it for profit — you don't just make a bad product. You corrode the thing itself.
I'll be honest. This is something I'm still wrestling with personally.
I build these things. I am a developer and an entrepreneur and a teacher simultaneously. Part of me wants to continue building, but I go back and forth between wanting access to be free and wanting to be compensated for my efforts and time. If I do it for financial gain, it will corrupt my intentions, and the work will no longer be to make education better and to help teachers and students in the truest sense. The profit will always be a factor in how I make changes and innovations. I don't know the answer.
But I think the fact that I can't resolve it cleanly is evidence that I understand what's at stake. The people who have resolved it — who decided education is a market and they're going to win it — they aren't wrestling anymore. And that's exactly the problem. The wrestling is the conscience. The moment you stop feeling the pull of both sides is the moment you've already chosen.
The history of AI in education is longer than most people realize. The dream of personalizing learning at scale is not new. It goes back a hundred years, to machines that quizzed students and moved forward only when they got it right. The technology has changed in every decade since. The ambition never did. What's new about generative AI isn't the goal. It's the scale and the believability.
But the question was never "will AI replace teachers?" That has always been the wrong question. The real question is what happens when the tool optimizes for the wrong thing. Engagement over understanding. Compliance over curiosity. Speed over depth. That's where educators have to hold the line. That's always been where educators have to hold the line.
We will figure out the right way through this. I believe that. Not because it will be easy, or because the people making the decisions will always get it right, but because the people closest to students — the ones in the room every day — have always found a way to do right by them even when the system didn't make it easy.
That's not optimism. Optimism is easy.
That's hope. And hope is what you choose when you've seen the full weight of the problem and you decide to keep going anyway.