Every time a new technology arrives, it finds its way into education, and every time, we are handed the same choice: listen to the people trying to sell it to us, or listen to teachers.

We’ve done this before. In the 1950s and 60s, television was going to be the answer. School districts wired classrooms for broadcast so a single “master teacher” could deliver a lesson to hundreds of classrooms at once. It was pitched exactly the way every new education technology gets pitched: as a way to reach more kids, more consistently, for less money. It didn’t hold up. A kid watching a screen is not the same as a kid in a room with an adult who knows them.

In 2012, Stanford’s then president John Hennessy, gave a keynote called “The Coming Tsunami in Educational Technology.” He predicted that technology was about to hit higher education the way it had already hit newspapers and music, and that colleges needed to drive down their own costs with online learning or get run over by whoever did. He pointed to Khan Academy, and to the first wave of MOOCs (Massive Open Online Courses offered to an unlimited number of learners worldwide) like Coursera and edX as proof it was already underway. He told the room to be the disruptor, not the disrupted.

“The tsunami is coming whether we like it or not,” Coursera founder Daphne Koller famously told The Atlantic. “You can be crushed or you can surf, and it is better to surf.”

That was over a decade ago. MOOCs never replaced universities. Completion rates on them were, and still are, famously terrible. The tsunami mostly reshaped corporate training, not the in-person college experience Hennessy was actually warning about.

Education has a problem the rest of the economy doesn’t. When a car company automates its assembly line, it builds more cars with fewer people, and the cost of the car goes down. A bank that used to need a room full of people to process transactions by hand, now does it in seconds with a computer. Technology makes those industries cheaper, and in some cases better for the end user. But, it has never done that in a classroom. What works in education hasn’t changed: one skilled, present teacher with a small group of kids. While pedagogy and curriculum can and do develop, nothing has ever made the core relationship of a caring adult working closely with a handful of kids cheaper without also making it worse. 

Still, that doesn’t stop people from trying to sell us the idea that it can. Every time a new technology shows up, the companies who stand to profit from it look at education and see a problem they can solve, because solving it would also make them economically successful, in a market worth a great deal of money. It’s really just Upton Sinclair’s old point. He wrote about efficiency being valued over humanity, looking hard at working conditions in the wake of the industrial revolution. He said it’s hard to get someone to see something clearly when their income depends on not seeing it. So the pitch comes fast, and every time, they tell us this one is different. It never is.

I have two stories from my own career where I chose the sales pitch over the people in the room who actually knew better, buying a company’s promise that a piece of technology would just make things better, and finding out that it didn’t. I think we’re about to make that same mistake with AI too.

Smart boards

If you don’t know what a smart board is, that’s because they were dumb. A smart board was a touchscreen panel mounted over a section of a classroom whiteboard, with a projector bolted above it throwing the image down. It ate up about seventy-five percent of your usable whiteboard space, and it worked maybe a quarter of the time.

At my last school, I went to a conference, got sold, and installed them in every classroom, over the objections of teachers who begged me not to take their whiteboard space away. The bill was over fifty thousand dollars. Within two years, almost every teacher in the building had quietly retreated to the small strip of real whiteboard they had left. I eventually took every one of them down and kept the projectors, which is, of course, what everyone eventually used in classrooms anyway: a whiteboard and a projector.

Much later in my career, I watched one of the most gifted middle school Waldorf science teachers draw by freehand in chalk  a cell in more depth and more beauty than I had ever seen in a textbook or on a slide. It made me question whether I needed a projector in the room at all.

Open gradebooks

The second story is harder, because the cost wasn’t financial, but human.

In 2008 when I was 8 years into my teaching career and a middle school advisor, the conversation about open gradebooks first started. Student management systems, the Ravenna or Google Classroom-type platforms, were becoming standard, and the question that came with them was whether to open the gradebook, making every assignment’s grade and every piece of feedback live to students and parents the moment it was entered. It was framed as transparency, and transparency sounds like an obviously good thing.

Shortly after, I became assistant dean and I was part of the team put in charge of rolling out open gradebooks.

The vendors sold it hard. And I remember exactly how the conversation shifted: it was never really a debate about whether we should do it. It became a conversation about when.

Every teacher I knew worried about the same thing. They said kids would start checking their grades constantly, obsessively, that a live grade/number would replace an actual understanding of how students were doing. They said teachers would stop assigning meaningful low-stakes work: early drafts and rough attempts that are supposed to count for little, so a kid can take risks and get real feedback that measures their growth before the pressure of a finished product. Once every point is visible in real time, that growth falls apart. Points become the only thing that matters, and learning stops being about growth and starts being about the running total. They said parents, once they had full-time visibility into a live number, would question grades constantly, and that teachers, already stretched thin, would get worn down by the volume of emails until grades quietly inflated just to keep the peace.

We did not listen to teachers’ concerns. We moved forward with open gradebooks. 

And the teachers’ concerns came true. Grades became inflated and, in turn, meaningless, and I’d argue that’s one of the biggest problems in education right now. Ask any college professor. It’s not just that they’re now fielding emails from parents complaining about their twenty-year-old’s grades, though that happens too. It’s that they’re teaching kids who show up with far more confidence than their actual ability level can back up.

The weight shifted onto tools that felt more objective from the outside: the SAT, the ACT, external standardized measures. That was already the direction college admissions were heading towards. Then Covid hit and paused it, test-optional policies spread fast. But it’s back. I saw it with seniors at my last school, and I see it now with Colorado seniors I work with: straight-A students earning SAT scores well below what straight A’s should predict, confused and frustrated that they’re not getting into the colleges they expected.

Denise Pope, who studies student stress at Stanford’s graduate school of education, has found that when you ask kids what stresses them out most, the answer is almost always the same: grades, tests, and other assessments. And that holds true across rural, urban, public, and private schools alike. I got to work with Denise, about ten years ago, when a friend of mine and I ran our own small version of that research. He was teaching in Boulder, I was in LA, and we pulled data from our own high schoolers on how often they checked their grades. The number came back between a staggering fifty and eighty times a day.

Because grades don’t mean anything anymore, feedback doesn’t either. Imagine if every piece of feedback from your boss was meaningless, and you never actually knew whether you were doing a good job or a bad one. That work environment would be toxic. You’d feel anxious – or apathetic – every single day.

Welcome to education in 2026.

The actual choice

This is the pattern, every time. On one side you have the people who stand to make money from  technology. On the other side you have teachers, who show up every day to do a job I genuinely believe is impossible. You never finish being a teacher. You just get used to trying as hard as you can, every day, for kids who are not yours.

I choose to listen to teachers on the question of technology in schools. In my experience, teachers have never been wrong. And on the issue of AI in schools, they are not divided. Almost every teacher I know is saying some version of the same thing: AI in schools, as it is currently being introduced, is not going to end well for this generation.

Not every piece of technology is useless in a classroom. A photocopier is a tool that makes it easier to prepare for class, which means more time to connect with students and build better curriculum. But it never replaces the teacher deciding what goes on the page. Something like Google Read&Write is the same kind of tool: it gives a kid with dyslexia a way to access the same text as everyone else and get their own ideas onto the page, without changing what the teacher is actually teaching or why. Neither tool replaces anyone. They just make the job a little easier.

AI could be that kind of tool. A teacher who used to spend hours building three or four versions of the same lesson for different learners could do it in minutes now, and spend the saved prep hours back on the kids instead. I’m genuinely open to that.

However, there’s a different version of AI on the table too. The real dividing line isn’t whether a teacher is the one standing up front delivering the lesson. It’s whether AI is expanding what a teacher can do, or replacing the judgment that makes them a teacher in the first place: reading the kid in front of them, deciding what this particular group needs, designing the experience that gets them there, and knowing when to change course. Some of what’s being sold right now does the second thing. It hands AI the decisions about what to teach and how, and asks the adults in the room to become something closer to coaches or monitors. It’s being sold as the future, the way everything is sold as the future. I look at that and think about every other time we believed we had found something that could replace a teacher’s judgment and their relationship with their kids, and every time, kids were the ones who paid for the experiment. I’m comfortable being behind the times while everyone else works out, again, that you cannot replace a teacher.

And this is why education is expensive. That’s the real problem underneath all of it: we have tried, over and over, to make education more efficient, and we can’t, because it never gets cheaper the way everything else does when technology advances. That’s why a school like Friends costs what it costs. It’s also why it’s worth it. Because a child’s education isn’t built on efficiency. It’s built on time, attention, core relationships, and the real people who know them and show up for them. Every single day.