Yeah the teaching training I got was "lead" them to the answer but don't give it to them.

If they get stuck. Guide them towards reflecting on the information they can use to get the answer.

Then they learn how to consider and reflect on knowledge to find conclusions.

Another big one is making it personally relevant.

"When did you experience or witness this in action?"

Or

"What's an analogy that you are familiar with that helps you understand this phenomenon?".

I can't see AI guiding students through something like that.

I was teaching a class and talking about changes in brain O2 use for a novice vs an expert at a task.

A student had mentioned in their introduction they played chess. So I used that example. My brain would use more oxygen to play chess (because I'm not very good or skilled) than his brain. But this higher activation of my brain isn't a sign of more brain power but the lack of skill. (I was illustrating how interpreting fMRI data isn't as simple as it looks since increases in activation (O2 use) isnt always a "good" thing).

Incorporating a student's experience into the example made it easier for all the students in the room to understand how it would work.

AI can't do that kind of thing either.