Andrea Jones-Rooy

Yes.

Well, and the other thing that is in what you just described goes back to your earlier point about attention, which is really not something that's in this study.

I mean, it's a very narrow study in that sense, in that it's like a toy kind of YouTube beginning, and you're following according to these rules and things like that.

But if you...

So there's sort of another ingredient where it's like there's the user making the choices.

There's the algorithm serving up videos.

But there's also the people who make the videos and decide on the thumbnails and the titles.

And that's sort of maybe the slot machine being engineered to catch your attention.

It could be the case.

And they did not do this analysis.

Another study should.

Maybe the words in...

particular ideologically extreme corners or echo chambers are more effectively calibrated titles of videos and thumbnails.

Like you've got to see this.

You won't believe who did that.

Like you're in danger for this reason.

And we have a ton of other research from other social media platforms

that shows that we're way more likely to click on videos or links or posts or whatever that invoke fear or anger.

And so it could be that part of the reason that people who are kind of racist curious end up very quickly becoming extreme is because the people creating the videos are being very deliberate in campaigning for your attention.

So that's another piece of it that really is left out of this study.