Andrea Jones-Rooy

But again, this could be rabbit holes about locksmiths or lawyers or whatever niche thing you happen to be looking up.

They find...

There is no evidence in their study of radicalization pathways where users on their own without user input, that just the pathway in the algorithm on its own, there is no evidence of radicalization pathways where users are driven into increasingly extreme content rabbit holes.

So it's without user interaction, it doesn't seem like there's something in the algorithm

pushing people to ideologically extreme content rabbit holes.

Finally, instead, the biggest thing they see as far as where politics comes in is that they say that YouTube pushes all users, regardless of their initial ideology, and they measure the ideology of the various people in their study, or they chose a certain mix of people from different ideologies.

So regardless of your ideology, YouTube pushes all users toward moderately conservative content.

and an increasingly narrow range of ideological content the longer they follow YouTube's algorithm.

So while it's the case that the echo chambers themselves are sort of mild, content rabbit holes are out there, but they could be about everything, fixing your sink, or it could be about politics.

There's no existing underlying radicalization pathway built in, but there is a general kind of nudge towards conservative content.

And once a user starts clicking into it, the ideological content, the rabbit hole does pick up and you can end up down there.

But it's the users, you know, it's kind of answering the question or starting to answer the question of like...

The user kind of has to have an interest in the extreme content in order for YouTube to push it.

And then once the user indicates interest, then you could kind of more quickly end up in that rabbit hole just as quickly as I end up in a cake decorating rabbit hole.

But I also thought it was interesting that there is a general push towards moderately conservative content.

Part of that might be because I believe, and I could be wrong, but I believe the evidence is suggesting that in terms of political content, there is more conservative than liberal content on YouTube right now.

So that could simply be a reflection of that.

But that's where we stand.

And I'm curious what you guys think.

I love these sorts of studies that say, hey, here's something we all kind of either take for granted or we argue about or we sort of assume is true.