It's Thursday, October. I'm Oscar Ramiras from the Daily Dive podcast in Los Angeles, and this is reopening America. The Facebook papers are out and it's shedding more light on how executives their weight, various trade offs between their bottom line and impacts on public safety. In one instance, it took them years to implement a fix for an algorithm that was feeding people angry emotional content filled with misinformation.
If you use the angry emoji on a post instead of the like button, it carried more weight, and then they would feed you more of the same despite the signal that you did not like it. Jeremy Merrill, Data report at the Washington Post joins us for more. Thanks for joining us, Jeremy, thanks for having me on it. Let's talk a little bit about the Facebook papers. We're learning a lot more about kind of how things were
working in the background at Facebook. You know a lot of different things, internal dialogues between between workers, and how things were being portrayed out there. One of the articles that you wrote about was about this, uh, these anger emojis, these new set of emojis that popped up in seen that lets people react to comments beyond the the just the like button. Um, you know, you can laugh at something,
heart something. You know, there's like a sad emoji or a sorry emoji I think they called it, uh, and then the the angry emoji also. But what happened with all of this is uh, things were waited differently, and these emojis were waited in such a way that it started promoting a lot of misinformation, a lot of things
that would give you more of an emotional response. And at the time they thought that this was the way to go, but it just kind of led to all sorts of mishaps and it took them a long time to really remedy that as well. So Jeremy, help us walk through some of this. What are we seeing with it?
That's exactly right. So what we're talking about here is whenever you go to Facebook dot com or open the Facebook app on your phone, you get your news feeds, right, and just just imagine all of the possible things that Facebook could show you. At the very top could be posts from your friends a couple of days ago. It could be a baby picture from someone you wrote the school bus with a high school fifteen years ago. It could be posts from a news organization or a buy
nothing group in your town. There's all sorts of things and Facebook has to make a decision of what you go first and what you goes back in what should go you know three thousands which that you'll never see it, And that algorithm is sort of one of the secret sauces that that keeps people, that keeps Facebook, um sort of a life. And broadly speaking, um, there are two
sets of things that make up that algorithm. One is a really complicated, really masthei for what you think of when you think of an algorithm that it's super complicated. That's meant to predict what's the likelihood that if we show this post to Jeremy, what's the likelihood that he's going to click like, what's the likelihood that he's going to click angry, what's the likelihood that he's going to
write a comment? Um? So that's the first part. Are all those predictions, and the second part is really simple. It's just numbers that are picks by human beings who worked in Facebook, and those numbers are how important relatively those different predictions are. And when Facebook launched a new way of ranking your news speed, a new way of picking what stuff shows up first. Those weights were that
alike is worth one point. These reaction emojis, including angry, but also all the other ones were worth five points, and comments were worth between fifteen and thirty points, and there's some other a couple other ingredients as well. But sort of the minute that this launched, someone from Facebook wrote on their internal message board and said, you know that they were worried that posts that make people react with the angry emoji might inadvertently open the door to spam, abuse, clickbait,
and somebody else said, you know it's possible. Yeah, I mean it makes natural sense, right, something that angers you, something that you're like, well, you know there that's full of crap. You know, I'm gonna anger emoji that hopefully you don't get any more of it. But it was actually the opposite, because they waited it so much more than just the like or whatever. Uh, the the algorithm would actually just serve you up more. And and in that sense, you know, the worst of the platform started, uh,
started shining through even more for people's news feeds. That's that's exactly right. If the algorithm learns that you're going to react annoying, it's worth five points, then it's going to show you more stuff like that, because that's what it considers. The Facebook's term was meaningful social interaction. They consider that to be a meaningful social interaction, so they're gonna try to do more of it. But gradually, um
evidence started building up. People found that things that sparked a lot of angry angry reactions were disproportionately misinformation what places Book called toxicity, as well as low quality news, and people started proposing, hey, maybe this isn't a great idea to count the anger emojis so high, and they counted as exactly the same as these other these other
emojis like love or care. Yeah, and that's part of what that's part of what's interesting about the Facebook papers and what we're kind of learning about how Facebook was operating is that while these experiments were going on, right how the different weights and how to serve up the algorithm best to people, the internal conversations from employees calling some of it out, saying, hey, this is a problem,
this is becoming a problem. This problem is too big, and uh, you know, really not being able to remedy it right away. Uh. That's one of the other interesting aspects about it, because that's part of what is We're we're seeing these internal dialogues happening behind the scenes, now right, That's that's exactly right, And you know, to be fair to Facebook, Um, in some cases people didn't have evidence
yet it was just a suspicious sort of an instinct. Hey, ranking anger so highly doesn't seem like a great idea, or it would be kind of weird, um to make distinctions between different kinds of emotions. They felt kind of uncomfortable with that, as well as raising the possibility that stuff that makes people angry angry might be important for social change. And um, but you know, eventually, this book did end up cutting the weight of this emoji all
the way down to zero. They ended up making distinctions between emotions. Um, So for a while, love and care were were double alike, and stuff that they maybe go ha ha or wow was only worth the same as alike.
So to their credit, they did end up making this fix eventually, But it was the case that they set these without They set the weights alto five times alike without looking into what would happen, And then once they've made that choice, they stuck with it for a couple of years until the weight of the evidence was so
much that they finally had to act. Yeah, and sticking with it too, right, And you know, it's a it's a new feature on the website, so it makes sense that they want to point people to it as much as possible. So that was part of the problem too. Is that why they didn't act so quickly, is that you know, they wanted people to start using that new feature, um thinking you know, it was going to be that next great thing for them. Uh, And so that posed
the problem too. And and as you mentioned once, they um did kind of start reducing the weights of these things. They found that, you know, in other parts of the algorithm also, you know, when a post would get too
popular or something, and they would demote certain things. Even then, it was it was tough because uh, you know, there was no cap on how many points let's say a post could get, so um, you know, even if you demote something, I mean, it still could have a billion points, right, and it could be at the top of everybody news feed. That's exactly right. It's really difficult for us to understand how all the sort of weird mass that goes into stay going out, what goes to the top of your
news feed. And some of the documents showed that even these people with access to a lot of resources that Facebook and a lot of people who can understand it offensibly, even they didn't really understand how the algorithm worked, and they were finding that, you know, the scores for the things that would appear at the top of your news speed usually are about a hundred, sorry, in the hundreds, a couple hundred, but some some of the post could
get scores in the millions up to a billion. And what what you've got is a sort of battle between uh, the integrity teams, who are the ones tasked with finding stuff that's probably bad and trying to push it down in people's feeds so it doesn't go viral, and the growth teams, who are tasked with getting people to use Facebook.
And so you should have had this battle where the growth team was giving these things these humongous scores, but the integrity team was limited and like the best that they could do is like cut it the score it half. So if you you know, the post number one has you know, five million as its score and post number two has eight hundred, you're trying to push post number one down because because it's probably garbage and all you can do is divide the five million and a half, well,
two and a half million is still huge. So they found that they had this trouble of actually making their principled ways of trying to make the platform safer and more enjoyable for people. They weren't able to work because the math was so complicated. They were always fighting that losing battle. And you know, it's it's interesting, you know, to the point, right, it's why is Facebook serving up so much crap? Why is it serving up so much anger and divisiveness? And this is exactly why it is.
The different weights that were put on things, and the math was so complicated it was hard to go back and fix it retroactively. So you know, we're learning a lot from these Facebook papers. Um, you know, Facebook choosing maximum engagement over user safety, dropping the ball in different areas. So we'll keep finding out more and uh and bring it all to you. Jeremy Merrill data reported at the Washington Post. Thank very much for joining us, Thanks for
having me on. I'm Oscar Roomiers and this has been reopening America. Don't forget effort. Today's big news stories you can check me out on the Daily Dive podcast every Monday through Friday, so follow us on I Heart Radio or wherever you get your podcast