How is AI going to change science?

Thanks, Roland.

And now to Caroline Steel.

Uh, how much have you got for us?

Well, I wanna be greedy and start with two papers.

So one that's been published recently, the other was published at the start of April, and they're both about geoengineering Arctic sea ice.

So as we all know, Arctic sea ice is melting at an alarming rate.

It's expected to be completely gone in summertime by the 2030s, which is bad news for polar bears, but also bad news for the whole planet because sea ice does this amazing thing where it basically reflects sunlight back into space and keeps the surface of the earth cooler.

So this geoengineering idea is basically to cover the surface of this ice in seawater, which would then hopefully freeze and then thicken the ice and slow down its melting.

That's the idea.

This is completely bonkers, isn't it?

[laughs] Well, it's definitely an out there idea, but it has now been, for the first time, tested actually by two different field trials, so one in Canada and one in Svalbard in Norway.

So scientists basically pump seawater from underneath the sea ice onto the surface, and both trials did successfully thicken the sea ice by, like, quite a substantial amount.

But this is the interesting or maybe slightly disappointing part.

So the scientists in Canada found that thickening the sea ice did slow the ice's melt, so it lasted for an extra 7 to 10 days in summer.

But the scientists in Norway basically didn't find that the ice lasted any longer.

It still melted by the same date.

So really mixed results.

But also we're talking about...

I mean, I don't know the sort of precise comparator, although it's normally safe to say the area of Western Europe 'cause you can change the size according to your needs.

But-

[laughs]

...

this is a very, very big area of ice-

Yeah

... that you'd have to do something to.