What Should I Teach in the Age of AI?

By Cha Seoyun

The more we can do with AI, the more I find myself wondering what I should teach.

Explaining things students don’t understand. Working through problems they struggle to solve alone. If AI can offer that help too, how will their expectations of a teacher change? I welcome it, but I don’t feel entirely at ease. The ways of teaching I’m used to may no longer be enough.

Still, I don’t want to assume that students who use AI can’t think for themselves. Sometimes we need help just to get started. Being able to make something that had only existed in your head, or ask a question without worrying about what others might think, is something to welcome. If we had to know everything before we began, many of us would never begin at all.

New technology excites me too. I’m not keeping up only because I feel obliged to learn. I wonder what I could do with it, and I want to try it myself. When I think of the pleasure students might find in it, I don’t want to discourage that curiosity.

But once we have a finished piece of work in front of us, asking what we learned becomes a little harder.

Getting a good piece of writing and being able to agree with the ideas in it are different things. Something may sound convincing at first, yet leave us stuck when someone asks why we think that way. Not because it is wrong, but because we haven’t thought it through ourselves. I suspect I, too, could readily accept an explanation I like simply because it fits what I already think.

So I’m more curious about how a student received an answer than whether they used AI. Which parts did they agree with? What didn’t make sense to them? Would they have accepted a different answer as well? I’m not necessarily looking for a neatly worded response. “Now that I think about it, I’m not sure” can give us something to talk about too.

I need to ask myself those questions as well. Being a teacher probably doesn’t make me any less susceptible to a convincing explanation.

I also want to be careful when I tell students to think for themselves.

If I’m going to give them time to work things out, I need to accept work that is still rough. If I tell them not to use AI, then expect their writing to be as polished as an AI response when I grade it, which expectation are they supposed to meet? It makes me wonder whether I ask them to use their own words while praising only conclusions that adults like to hear.

I would probably be tempted to give higher marks to a polished result. It is easier to read, leaves fewer gaps, and explains things clearly. But how should I read a piece that is a little awkward yet shows that its writer has spent a long time thinking? If we’re going to talk about education in the age of AI, I think we need to face uncomfortable questions like that too.

That is why teaching a new tool isn’t enough. My own judgment of what counts as good work, and which efforts deserve recognition, may need to change. I don’t yet know how far.

I’ve also been thinking about the pleasure of learning.

As soon as something interests us, we often ask what we can use it for or what results it will produce. That becomes even more common as we grow older. It takes time and money, so of course we consider its usefulness. Still, it would be exhausting to have to justify everything we enjoy as something that will help us later.

I might tell a student to go ahead and try, then hesitate to do the same myself because it feels like a waste of time. Teachers need time to be clumsy learners too, but I feel I ought to know things before my students do. It would be a shame to avoid something new because I don’t want anyone to see what I don’t know.

On the other hand, I don’t want to spend all my time for rest on learning, either. If I have to keep up with every new technology, there will be no end to it. I’m not confident I can always distinguish what I need now from what simply interests me, or from what I’m looking into because everyone else is doing it and I feel anxious. At least I hope I don’t treat all three as the same reason.

I agree that we need to teach students to use AI well. I just hope “well” means more than how quickly they work or how much they produce.

I also want to hear what students want to do. With more things becoming easy to do, choosing may actually get harder. We don’t have to do everything simply because we can, and we don’t have to enjoy the same things as other people just because they have achieved something with them. That is something I want to tell my students, and something I need to hear myself.

What we need to know will probably keep changing. Will adding new material each time be enough? Among the things I have considered important until now, are there some I could teach less of?

I’ve often thought about what more I should teach. I’m not sure how much I’ve thought about what less I should ask of my students.