One thing we know about the future of education is that it won’t—or shouldn’t—look like rows of students sitting at desks facing a teacher at the front of the room. Even if the teacher is, or is accompanied by, a robot. Even if the robot has glowing eyes that put it only a colour shift away from Terminator territory.

What we don’t know about the future of education at the moment would fill an encyclopedia. And this is an uncomfortable place to be for educators who are used to, comfortable in, the role of oracle, providing wisdom to students seeking to find their way in the world. How can you give them the guidance and security they need if you don’t know what they’re going to have to deal with in the world they’ll shortly enter?

This is one of the many daunting dilemmas that have been forced on teachers by the exponential march of artificial intelligence. Like most of them, this one resolves through confronting an uncomfortable dichotomy: Realizing that predictability is no longer available to us, but that isn’t what our students need the most anyway.

Yes, we’ve been saying “Guide on the side, not Sage on the stage” for a long time, but easier said than lived, right? But now AI gives us no choice. It has yanked the rug from under us and put us in a position where often our students know more than we do about this technology—and far from being a frivolous smartphone app like Snapchat, this one sits astride the dual potential outcomes of vastly accelerated learning or cognitive atrophy.¹

My work since 2016 has been to give people agency over AI in their lives; to empower them to choose to face important questions before AI forces those questions upon them. The most significant of these questions is “What does it mean to be human?”, but we can refine that into one more incisive for our time:

Who do you want to be in a world where machines can think?

This simple question sets so much in motion that in my work with schools this is our starting point, our North Star. Whenever I’m at a loss for a strategy, I find it restores clarity and direction.

And it comes to our rescue when we’re struggling with that basic question of What value can I provide students when I don’t know the technology, don’t know how the workplace is changing, don’t know how my students will function in tomorrow’s workplaces or whether they will even have jobs—or whether I will, too? (If you’ve not experienced that existential doubt, then you haven’t fully realized or confronted what AI means. And you need to, because your students are going to meet that doubt head-on, if they haven’t already, and you will lose rapport if you don’t know what they’re going through.)

Our North Star points us to something more valuable than predictability. It led me to a realization that I’ve used to help people who have come to me in anxiety about the future—if any—of their jobs. We’re used to a triad that goes Prediction—Planning—Preparation:

• First I predict what’s going to happen
• Then I can make plans for how to deal with it
• Then I can prepare to execute those plans

All well and good, but I submit that for our new era, those are exactly backwards. And we can see why if we describe those stages in terms of their future focus:

• Prediction is a future focus on What will the world do?
• Planning is a future focus on What will I do?
• Preparation is a future focus on Who will I be?

Now, which of those gives you the most agency? Which gives you the most control, the least dependence on external factors?

It’s obvious, of course. But we get so hung up in attempting to predict everything that we think that we can’t function without those predictions, and we seek the best pundits for their predictions. When those pundits organize behind a brand, we call them a consultancy.

Of course we need some predictions. But less than most people think. What’s our greatest fear, generically? The usual answer boils down to The Unknown. This is where the monsters under the bed derive their power.

But there’s more to it than that. Consider: Have you ever done anything like backpacking through a wilderness, jungle, desert, etc that you hadn’t been to before? A trip to the Amazon, outback, or similar? Were you looking forward to it? Wasn’t it precisely the fact that you were going into the unknown that was exciting?

The difference is agency. You were making the choice, instead of being kidnapped and dropped naked into the backwoods to fend for yourself.

Our North Star question sets the direction. The world is changing—yes. Our students are launching into that world—yes. But rather than asking how they will fit into whatever that world is, ask how they will shape it. What is important to them? How will they make that happen? Who do they need to be to have that agency?

And who do you need to be to teach them?

Of course they need skills and knowledge for the workplace, but think first about what you have to give them that they can’t get from AI. (Believe me, when ChatGPT came out, I woke up every day asking myself, “What can I give an audience that they can’t get from ChatGPT?” Because I couldn’t face them until I knew that I was not wasting their time.)

And one thing I can guarantee is, no student will remember something ChatGPT told them for the rest of their lives. But they may remember something that you, the teacher, told them. Where you find the difference between you and AI is where the human race starts answering the same North Star question.

What’s the role of AI in the classroom?

The North Star points us in the right direction but it doesn’t take us all the way. “That’s great, but I’ve still got AP Geography tomorrow morning; what am I supposed to do?”

If that question had an answer that would fit on this page there would be no need for this blog and no need for our organization. The fact that we don’t yet have succinct answers to how the classroom process changes shows that those answers are intimately bound up in who we are or choose to be as teachers and schools. Which is why our work is so deeply interactive and experiential.

Many answers have been given that are maddeningly off-target. One bubbly suggestion involves embracing AI as much as possible, and ask the students to:

• Write prompts for their assignments and grade the AI’s outputs;
• Use AI to generate images and videos accomplishing the learning outcomes;
• Use AI to create infographics and storyboards explaining the topic;

It’s not that these can’t be engaging activities; it’s the question of What are these activities designed to teach?

• How to write a prompt?
• How to create images and videos with AI (and correct their inevitable mistakes and misinterpretations)?
• How to format an infographic (and either persuade the AI to spell correctly or battle with presentation tools to learn layout and font choices)?
• How to critically evaluate potential answers, vet sources, and identify missing crucial points?

All these are useful learning opportunities, but not when they’re convolved into a single assignment in AP Geography. How are you going to decide what to assess out of all that?

We need independent training in AI literacy for students. You wouldn’t expect the Grade 9 Physics teacher to be educating students in elementary grammar so the students can produce readable answers to their dynamics assignments—well, you wouldn’t want to, right? So why should every teacher be embarking on their own journey of educating the same students in the same skills of using AI without coordination, cooperation, or collaboration? That’s a colossal waste.

And it’s a colossal waste if we don’t take advantage of what AI now provides. A couple of years ago, a professor at a large university called me for advice, and one of the things I said was, “Imagine that I gave each of your students an expert tutor in every subject, free of charge, available 24/7. How would you adapt your teaching methods to leverage that?” And he said, “Wow, that would be incredible, I’ve never thought about such a thing, I have no idea what I’d do.” And I said—because this was March 2023, 4 months after ChatGPT came out—”But that’s what you already have right now.“

And no teacher would turn down the offer of a private tutor being made available to their students. Unfortunately the AI versions of these lack integrity and if you ask them “Please would you just do my homework for me?” they’ll say, “Sure, gimme the keyboard.”

Our job is to set up our systems so that the students don’t want to ask the AIs to do that.

That sounds impossible, and in many cases, it may be. What we’re seeing right now as the way of holding the line on academic integrity is for teachers to get to know their students’ writing intimately—perhaps through some initial proctored classroom assignment—because while studies show that teachers cannot reliably tell whether a piece of writing was done by a human or an AI (and AI detection programs do not work), they can tell whether a piece of writing was done by a particular student they’ve gotten to know. The teachers who have adopted this strategy are generally satisfied with that choice.

If your student and workload ratios don’t permit gaining that kind of familiarity, your educational model is in peril.

So our classroom of the future will look less like students arrayed in subordinate rows, perpetuating hierarchies, but more collaborative. Give them the chance and they will show you things AI can do that you never thought to try. They’ve grown up with it, whereas we have decades of experience of what computers can’t do to hold us back from thinking of new possibilities. Put guardrails around them to keep them safe while they innovate. They’ll invent the future in front of your eyes, and you’ll model for them the character of someone who faces the unknown future with excitement.

1. Yes, I’ve got em-dashes in the text. No, none of this was written by AI. I’m a writer, I know when to use em-dashes, and I’m not going to stop using them just because AI has overused them. Besides, the latest LLMs have largely gotten over their em-dash fetish. ↩︎