Eileen

And is that leading to

to certain kinds of holes or differences in their development, which are problematic.

And I would argue that instances of both are existing in our use of AI right now.

And we need really to get to the bottom of how we want our young children to develop in today's world.

And that should be the driving north star toward which we point our designs that are going to leverage AI.

That's exactly right.

So we really need to think back and look and say, what is the goal of this learning exercise, actually?

Is it to move your pencil on the paper?

Sometimes it actually may be when you're five or you're six years old, right?

And that you're actually learning motor skill and planning skill by learning to handwrite.

It doesn't matter if you can type later.

The act of learning to move your hand in these planful ways turns out to have important developmental affordances for the brain.

Or is this a skill that we really don't need anymore, that modern society can skip over because it's something that doesn't seem to actually develop dispositions or propensities of mind that are going to be useful later and we can drop it and let it go.

Yes, well, Chris, you're absolutely right that these ideas that I'm bringing forth now echo and build from and incorporate many kinds of pedagogical approaches that have been developed over ages.

And so whereas I won't speak to any specific educational approach like Montessori versus others,

These approaches vary in that they are focusing young people on the opportunities to make choices and to agentically move through the learning opportunity.

The ways in which those learning opportunities are set up will have very important implications for the ways in which young people learn.

For example, together with a young scientist named Solange Denervaux, who is in Geneva, Switzerland, we did some studies of the effect of Montessori classroom experience versus traditional classroom experience, and we wanted to understand whether those experiences of learning the math would actually shape the way the kids processed math in real time in an MRI scanner.

So what we did, Solange had 8 to 12-year-old math students from Montessori schools and traditional schools in Geneva come to her lab and solve math problems in the scanner, among other things, while we were scanning their brain.

And the first thing that we found was that the Montessori schooled kids in this context and the traditionally schooled kids got the same number of math problems right.