I’ve been chatting to colleagues and friends in academia and industry about AI, and a lot of us share a sense of gratitude that we studied and then worked before AI. We have more than a decade of experience doing research or working without AI. And we feel that, even if we embrace AI now, this gives us a leg up because we have this foundation to rely on. We’ve done the things by hand, so we know our stuff, and can now even do our work faster and better.

I’ve also been chatting to some of my postgraduate students, and they share a similar sentiment. They are grateful that they completed their undergraduate degrees before AI really took off. This has given them a firm underpinning for the fundamentals. They really know their stuff. And, because of that, they can now use AI wisely for their postgraduate work. They know when to use it for mundane things, and they know when they should not outsource their learning and rather do things by hand.

Then I had the most fascinating discussion with a first-year undergraduate engineering student. He told me how he is really concerned about kids in high school today. He is so grateful that he completed high school without AI, because this gave him a firm foundation for the fundamentals. It allows him to now use AI wisely. He also feels that we should urgently invest in teaching high school kids how to use AI properly. They are lost on their own.

I hear this idea that we should “teach AI” from the top all the way down: Many of my colleagues say we should introduce new postgraduate courses on how to use AI, my postgraduate students say we should introduce undergraduate courses on how to use AI, and, apparently, my undergraduate students say we should introduce high school courses on how to use AI. We all think the layer below us is lost – and that just can’t be true.

I am not saying we shouldn’t think carefully about teaching AI, but this first-year student’s comments really shook my thinking on this. (I am still trying to figure out exactly what I should take away though.) The one thing I do know: when I have a question about how to use AI properly, who do I ask? My line manager? The course developers at our university’s department for teaching and learning? No. I ask my postgraduate students what they are doing. I ask the layer below me. Maybe I should go and consult some high school kids as well.