Joel Pearson

How can you introduce it?

And AI is like, there's a lot of these things, like there's the bad and the good.

And how do you figure out a path to balance those two things?

I would sort of recommend bringing it in a way that is least addictive, least stimulating, and is a way that is project-based and can help them solve something they're working on.

Yeah, I mean, it's hard to put a – I don't like putting a hard number on that.

Yeah, early teens or before.

There's not, I mean, there's some learning ones, like the Department of Education has one in Australia.

Khan Academy made this Khan Amigo.

So there's educational ones, which are designed for young people, for kids.

But that's different to just jumping on one of these large language models and just going to town.

But yeah, I mean, because they're, at least for the US-based AI companies, they're racing to get to the point where their AIs can build themselves and improve themselves.

And that's their goal.

So mostly they're not interested in how to help society at the moment or how to help society integrate or how to help businesses use AI right or how to build AIs for kids yet.

That will come later, I think.

But right now they're all just in this sort of arms race to who can get to this self-improvement and this sort of super AI thing.

These AGI terms I'm not a fan of because they kind of are meaningless.

But once AI can go and improve itself without any human interaction, then that's this threshold where we'll just hit the vertical and the exponential.

Yeah.

And I'll often say that, particularly when talking to businesses who just think of AI as another piece of software, they'll buy a license and go and give it to their staff and say, you know, go and use this now kind of thing and force it upon them.

And that does not prepare, that just doesn't work.