SIO258: How Social Media Can Subvert Democracy

Transcript

This transcript was generated automatically and may contain errors — the audio is the record.

0:00:29 Thomas: Hello, and welcome to Serious Inquiries Only, this is episode 258, I'm Thomas, and that's Lindsey Osterman, back at it again, how's going, Lindsey? I know, yeah, just everything's perfect, nothing, nothing really to talk about, it's all, for all the people in the future listening to this, everything's just fine, nothing wrong, great day.

0:00:37 Lindsey: Oh, fantastic, couldn't be better, uh-huh, yeah, slow news month, yeah.

0:00:51 Thomas: It's uh, by the way, it's October 14th, uh, and it's 93 degrees where I am, like, does it ever? It's just summer all year round and like, I know people must, some places must think that sounds amazing, but I'm tired of it because like, it's not as though not it stays 93, like, it means like, you know, 108 most of the summer. And then like, like, eventually it gets down to the mere ninety five, ninety four. And then by like November, maybe it'll be what, eighty five. I don't know. Like, does it ever end?

0:00:57 Lindsey: Oh, no, no I used to live in Mississippi, so it was, it was pretty awesome there too, yeah , yeah,

0:01:27 Thomas: That's right. It's dry heat, though. It'S dry Heat here. So it's all fine. Now, I know it is better. It is way better than humidity, but still, it still sucks when your car melts onto a puddle of, you know, molten Terminator 2 fluid on the road. You know, it doesn't matter that it's a dry heat, but that's all

0:01:35 Lindsey: Yeah it sucks, you're, you bringing back fond memories, I gotta say, yeah.

0:01:48 Thomas: Right. Well, here's what we're here to do today. So you might have gotten the impression we kind of both realized based on the Social Dilemma episodes, which were great. Check them out if you haven't heard them talking about some of the science behind the claims in the Social Dilemma documentary on Netflix. It was, you know, maybe not entirely accurate in some places. We did that and, you know, we're thinking about it after the fact. I think both of us were thinking, you know, given that we kind of emphasized and talked about some of the places where the doc went wrong, potentially, we didn't want to leave people with the impression that like, okay, so therefore all social media is hunky-dory, everything's okay. And I think we were careful to say that at the time, but, you know, it's worth going into. I think we both thought it would be worth going into some of the darker sides of social media. Stuff that maybe should have been covered more in that documentary, don't you think?

0:01:48 Lindsey: Yeah, oh, Yeah, absolutely. I mean, so those were my two major claims, it's like they spent all this time on these claims that were not very well supported, and then spent what, like, all of 15 minutes on, you know, genocide. Yeah.

0:02:57 Thomas: Yeah. Yeah, a couple minutes in genocide and that's it. They had to make room for the lovely reenactments with madman actors and stuff like that. So, you know. Couldn't squeeze it in. So, we're gonna do it here. Here, maybe, maybe we were making you a little less freaked out last episode than you might have been. Now, maybe we're making you more freaked out. We'll see.

0:03:01 Lindsey: Yeah.

0:03:14 Lindsey: Hopefully

0:03:23 Thomas: We'll See what Lindsay has to say.

0:03:21 Lindsey: Yeah, if You Don't Come Away From this freaked out, you're, you're not, uh, you're Nicht listening hard enough.

0:03:26 Thomas: Ooh, okay. Well, that being said, why don't you take us through the, uh, the science, the findings

0:03:27 Lindsey: Yeah. So the, the second claim here, just to frame it, is that, yeah, there are really problematic ways that we should all be concerned about and paying attention to, in which, you know, social media is being used as a tool for nefarious ends. And so what I thought would be fun to do a deep dive in today is to talk about some well-established ways in which we are imperfect perceivers of our worlds, and particularly our social worlds.

0:03:57 Thomas: You don't say.

0:03:58 Lindsey: I know. And the ways in in which social media amplifies those imperfections. And they mostly do this by accident, but they do it nonetheless. And then finally, the ways in which social media can be used intentionally to exploit those imperfection to mobilize action, which can be a good thing if your intentions are good. But it is actively being used for, again, nefarious purposes in a lot of areas of the world where there's no oversight because it doesn't affect Zuckerberg's bottom line.

0:04:31 Thomas: Yeah.

0:04:32 Lindsey: So, yeah, so that's where we're going. All right, so let's start with these cognitive biases. So, again, like the overall claim is that, you know, we feel oftentimes like we are sort of objective, passive perceivers of our worlds, and we're definitely not. And so a lot of these are probably going to seem familiar to people to some extent, at least. So confirmation biases, everybody knows about those, right? This is the basic observation that we tend to interact with information in a way that confirms our existing beliefs. And so there's a number of ways that this works. So for example, we're more likely to seek out information preferentially that's consistent with what we already believe. Got a very accessible example of this that happened to me recently. I got a fever and a cough.

0:04:32 Thomas: And the only prescription is less cognitive biases. No.

0:05:24 Lindsey: Exactly. Yeah, fever, cough and fatigue. And so I had a pretty strong belief about why I had those symptoms. Pretty sure it was COVID. And so as one does, I googled. So what else are you going to do? Aside from go to the doctor, like a responsible person. But, you know, I had two strategies in front of me, right? I could Google my symptoms and sort of start with the evidence in order to arrive at some hypotheses about what's wrong with me. Or I can start with my hypothesis.

0:05:27 Thomas: Yeah. Right. Right. Yeah.

0:05:55 Thomas: Gotcha. You go the other way and then you match your symptoms and then voila.

0:05:57 Lindsey: And yeah, exactly.

0:06:01 Thomas: I'll say, though, it's not the... I mean, doesn't seem like the craziest thing to be worried about the fact that you might have the illness that is a pandemic that's currently happening. That doesn't strike me as that cognitively irresponsible.

0:06:13 Lindsey: No, no, no. I think in this case, it was reasonable for me to get a COVID test, for sure. But I do think that but it illustrates the bias fairly well, right? Like our impulse is generally not to start from the bottom up. We tend to start with our conclusions and then use that to guide our search.

0:06:20 Thomas: Yeah, right. Right.

0:06:36 Lindsey: So, you know, I Googled COVID symptoms and Google gave me a nice list of all of my symptoms plus, you know, a few other ones. Right. And so I was like, all right, there we go. I got COVID. Okay. So that's, that's one mechanism through which confirmation bias works. Another, another, ooh,

0:06:44 Thomas: Yeah. Yeah, my version of that happened yesterday when I'm so glad I would advise everybody if you can, if you find that they're not expensive to buy one of those oximeter things, not expensive. If there was like, I want to say 25 bucks. Who knows if somebody's hoarding them or something at this point. But I didn't buy it early on. I bought it only a few months back. But anyway, because my big fear was that like I would not realize that I had something. Something. And then, you know, you hear these stories about like, well, all of a sudden they, the person's blood oxygen was down.

They didn't even realize it and blah, blah, like all this stuff. And I was like, Well, can I buy one? So I looked at, looked them up, not very expensive, very simple, tiny piece of machinery, the thing that goes on your finger, you know? And I was like, I'm going to buy one of these. Cause if ever I'm freaking out like an idiot, cause I think I'm dying for no reason. I mean, again, not no reason, cause we're, we are in a pandemic, so it's got to give ourselves a little bit of a break here. But then I'll just pop this thing on and then that'll very much help me feel better.

And if I got one, by the way, that was just purely like placebo and just showed like 98 every time, that'd be so funny. But like that would probably be better for me in the long run. No, but so yesterday I decided to do some exercise for the first time in eight months because I have to accept that I can't just hold out until hockey comes back and the pandemic's over. Cause I'll just be dead probably from sitting around. And so I did some exercise and my lungs, oh my God, they did not feel good.

0:07:02 Lindsey: Yeah, yeah.

0:08:23 Thomas: Like, yeah.

0:08:23 Lindsey: I know, it's a rude awakening.

0:08:24 Thomas: And I was in shape eight months ago. Uh, and it felt, but it felt bad. And I, I think I've had a bit of a cold. I'm not sure, you know, and it fell bad to the point where I was like, Oh, I'm dying. It's COVID. I have it. I'm already dead. And then I, after the workout, you know, like I'm still like, can't breathe very well. So pop on that little thingy on my finger, 98. I was like, okay, I'm fine. That was my version of it.

0:08:36 Lindsey: Yep, definitely got COVID. Mm-hmm.

0:08:48 Lindsey: God, I gotta get one of those. Sounds awesome. It seems like that'd be helpful, I mean, because another fun thing about COVID is that the symptoms mimic a panic attack, or maybe I should say that in reverse. Yeah, that's

0:08:58 Thomas: Yeah. Oh, that's everything. If you're someone who has anxiety, like I kind of do and all of my family does basically like we can will ourselves into any illness. If we tried hard enough, we read enough about it. We would have that illness by the time we finished reading about it, I

0:09:02 Lindsey: True. Mm hmm. Yep. Yeah. Yeah, but I mean it was, it was insidious. It was like, I, I, that actually happened, I think, pretty shortly after the shutdown that I, like, started uh being hyper vigilant about my breathing and then I was like, oh nope, this is definitely shortness of breath. And it just escalated and I just had a panic attack for an hour. It was awesome, um, but like several of my friends had that same experience, so I don't know, maybe we should all get oximeters. Sounds

0:09:36 Thomas: Also I can just pop them on my kids. Works fine on the kids too. Make sure they're not dying. Cause how do I know they're not dying? I don't know. Anyway, it's, it helps.

0:09:42 Lindsey: We should just all have those is permanently installed. I think that'd be great.

0:09:45 Thomas: Yeah. So, so okay. Cognitive biases starting from

0:09:49 Lindsey: Cognitive biases.

0:09:49 Thomas: The wrong way around. That was a good example.

0:09:52 Lindsey: Yeah. So second mechanism is that, so we tend to evaluate information differently based on whether we agree with it or not. So even if we, even if I can resist that first impulse to start with my conclusion in Googling, I still have this to get over too. So let's say that instead of Googling COVID symptoms, I instead just typed in my symptoms to see what diagnoses come up, which I actually did after I did the first thing. And of course, COVID came up because everybody's Googling this right now. So that wasn't helpful.

0:10:19 Thomas: Well, and so you're like, now I definitely have it. Cause both ways I arrived at the same conclusion.

0:10:24 Lindsey: Yeah. I tried to be a scientist and I got The same, same inclusion. So yeah, but so COVID came up, but also, you know, the common cold and the flu and allergies all came up as well. And still, I had, I noticed myself doing this thing where I was evaluating the symptom list differently. So I was sort of paying more attention to the things that were consistent with the COVID diagnosis and disregarding the symptoms that I didn't have. And simultaneously, I was putting more emphasis on the inconsistent symptoms for the other three diagnoses. So with a cold, you usually have nasal congestion. I didn's have that. And so I was like, well, it can't be a cold. That's right out. But I also didn't lose my sense of taste or smell, and that didn't rule out COVID for me. So I was evaluating that information differently.

0:10:27 Thomas: Yeah, right.

0:11:14 Lindsey: And I think that another accessible example right now, and this is closer to where we're going with social media, is like the way that we evaluate news sources. Like I have a very different sort of eye when I'm reading something from Fox News as opposed to when I'm reading something from the New York Times. You know, I try to be critical in general, but, you know, I feel like I can relax a little bit if I'm reading a New York Times article, assume that they've checked their facts and things. And for something on Fox News, I'm usually sort of reading specifically for the things they might have got wrong. I'm probably doing a lot more active fact checking, that kind of thing. And so this is consistent with, you know, a large body of research indicating that people do this in general. If you encounter an argument that's inconsistent with what you already believe, you're likely to be a lot More critical about it, essentially.

0:12:02 Lindsey: And then the last thing is that that we tend to create information that's consistent with what we already believe. And so this is referring to the fact that we are not passive observers of our social environments. We are active observers. We participate in them. And so for this one, you can think like self-fulfilling prophecies. So if I have a belief that Trump supporters are assholes, that might lead me to be a little snarkier with them when I'm interacting with them on social media, and that might provoke them into being assholes. And so that would be a way that I'm interacting with the world in a way that creates information that confirms.

0:12:36 Thomas: I feel like a whole lot of that has has happened going, going both ways.

0:12:40 Lindsey: You think? Yeah, I think you might be right about that, right? Okay, so obviously social media didn't create any of these things. Confirmation biases have been established for a really long time. But I think that media is amplifying them in some fairly obvious ways, right? So like, know, if we think about algorithms on YouTube, right, we talked about that a little bit last time, the fact that these are set up so that they are suggesting videos to you that are similar to the videos that you've already watched is sort of feeding directly into this tendency to seek out things that are consistent with our pre-existing beliefs,

0:12:41 Thomas: Yeah. Yeah, but somehow it also gets worse, which I'm sure you're going to talk about. Like, it's not just like it shows you similar things. It's like it somehow shows you sequentially worse things until you believe the moon landing's picked.

0:13:22 Lindsey: Mm-hmm.

0:13:31 Lindsey: Yeah, exactly. Right. And I mean, yeah, they're doing that because, oh gosh. So again, I've got to plug rabbit hole again. They talk specifically about why they increase the sort of extremity or intensity of those recommendations over time.

0:13:46 Thomas: Yeah. Allegedly they've fixed that, right? Like, okay. So maybe that's fixed. I'm sure everyone breathes a sigh of relief.

0:13:49 Lindsey: Yeah. Yeah? Yeah. Uh-huh. Yeah but there's a lot of ways in which it's not, right? I mean, you can see this in the way that our ads are targeted. We're to see ads for things that are consistent with things that we've shopped for or pages that we'd liked or whatever it is. And also, the algorithms that sort of regulate our newsfeeds, right, we're more likely to see posts from friends that we interact with a lot more, which is a good proxy for how much we like them and oh, boy.

0:13:58 Thomas: Yeah. Okay.

0:14:22 Thomas: Apologies. Now I got to tell another stupid Thomas thing again. Because I had a gentleman, and you already would know this gentleman just by the description, I'm sure, was posting some really, really Bernie left anti-Democrat stuff and about how they're just as bad, blah, blah. All that kind of crap. You know, and I, I don't, you know, I'm friends with a lot of people through the podcast. I don't know who this person is, but I saw the thing. And on one thing, I think two things in a row, sometimes I get in a mood, you know, we all do our stupid stuff on social media. Sometimes I get into a mood where I just want to like break their echo chamber slightly with all the people who are like, yeah, and Biden where, and so they post some conspiracy theory about how like Democrats are just trying to get Trump to win because then they can stay in power in the Senate. I was like, that's stupid. I was just like, there was a couple of links in a row. I was like, that's stupid. That's stupid, I didn't care. That was just for fun.

0:14:32 Lindsey: Uh-huh.

0:15:18 Thomas: I felt like doing that. And so I think he responded and I was, I, I didn't even care. I was I'm not going to, I just was being, kind of dumb, but I felt at least voicing my opinion that this is a stupid conspiracy theory. And so a couple that went by and then Facebook starts showing me all this guy's stuff. And I was like, okay, hide, hide. And I, didn't, I full on block the guy or hide the guy. Cause I usually don't want to do that.

You know, like I would rather just, okay, I'll still be friends with someone if they, you know, whatever, whatever reason. But I, to me, that feels like that's a decision that I have a threshold for that. I wasn't across that fresh threshold. I wasn't like, okay. I want to block this pretty well. I, in my perfect mind, how I want this technology to work is I don't really want to see that guy's stuff much unless, you know somehow he has a post that a bunch of people are commenting on or something, you know, like that's how the algorithm should work in my mind. Like, unless this guy has

0:15:34 Lindsey: Right. Yeah.

0:16:09 Thomas: An important post in some ways, like something that's made a lot of waves, then, then I should see it. And I hid, so I didn't want to hide him entirely, even though I should have, but I just hid, I think two or three posts. Cause I kept seeing, I was like, hide this, hide this, hide this post, hide this post. And it still shows him to me every day. Now I see that guy and like, so what's going on there?

0:16:09 Lindsey: I Don't know, I mean, you know, maybe Facebook has changed their algorithms a little bit in response to some of these criticisms, because I'm definitely not the first person to point out that this is problematic, I mean, yeah.

0:16:41 Thomas: Yeah. But I guess because I commented on it once that makes it, I don't know if, if they're doing the, the really sinister madman guy thing of like, Oh, he was mad at that guy. So let's keep it, you know, that kind of thing. I don't know if I would. Yeah. Like that's, I don't know if I believe that's what they're doing, but just the fact that I commented once means despite hiding two or three more after that, it still thinks I have to see this guy's posts.

0:16:48 Lindsey: No. No. Yeah.

0:17:05 Lindsey: It's it's That last thing that makes me think that maybe they've changed it because, you know, you interacting with him and that prompting you to see more of his stuff, that makes sense. That's consistent with the stuff that we're talking about. But yeah, you would think that hiding would also feed into the algorithm.

0:17:16 Thomas: Sure. Apparently not.

0:17:22 Lindsey: Well, maybe, because I think that some of these things are reciprocal. I'm not 100% sure of that, but

0:17:31 Thomas: He chose like a, I want to annoy Thomas option on his feed or something.

0:17:36 Lindsey: No, no, but like maybe, maybe he's interacting with your stuff more and you're not noticing it or something and that's

0:17:42 Thomas: Oh, yeah, I haven't noticed that. But I don't know, like, you know, I'm kind of a big deal. No, I'm just kidding. We have a lot through the podcast. There's plenty of people at all times interacting with my stuff, you know, and like, I don't know, you know, it seems that's what's weird is I have, you know, a couple thousand friends who are listeners or whatever. And I seem to see the same like 10 people's stuff somehow all the time. Interesting. Anyway.

0:17:43 Lindsey: Yeah. It's Right. I mean, at least in part and maybe in large part it's, it's, it's because of this, right? It's like it's a function of of who you tend to interact with and things like that, um, and messages probably I think feeds into it

0:18:14 Thomas: Yeah. Right. But isn't that just going to perpetuate itself. I only interact with the stuff I see, and so

0:18:22 Lindsey: Yeah. Yeah. It does. No, that, dude, that's the whole point. Yeah. It perpetuates itself, um,

0:18:25 Thomas: Yeah. Yeah. Yeah. Exactly. And, you know, Facebook should have a thing where like, you know, someone you haven't interacted with in 10 years, it's like let's just show them that and just in case. Yeah.

0:18:31 Lindsey: Yeah. And, you know, for all we know, they are trying out things like that to see how it affects their bottom line. And so if that ends up being a good thing, then great, but they're definitely not doing it to make your perception of reality more accurate. That's, that's not in their best interest, um, and that's the whole problem. I actually, so there is one thing I want to, I want to sort of pick up on in what you were saying, um, the idea that like in a, in a perfect world, what social media would do is like, um, yeah, yeah, mostly show you things that are from people that you like and agree with or whatever, but like occasionally show you these other things that seems very parallel to what we actually do in the offline world, right?

Like we're, we're cure, we always have been curating our social worlds. We've, we've always been in the position of wanting to associate more and hang out more with people that we agree with. But then we also have these, these situations at work, or maybe just, you know, just incidentally interacting with people out in public, or we used to before COVID, um, where we, we sort of interface peripherally with people that we don't agree with.

0:18:46 Thomas: Right. Yeah.

0:19:41 Lindsey: And so I think that, yeah, ideally we would, we would have a system that set up a little bit more like that. Um, so that we're sort of getting that exposure more, but unfortunately that's, that's not the way these work because of misaligned incentives. Alright. Yeah. We've got these mechanisms in the confirmation bias that to the self-reinforcing cycle. We believe things, and that leads us to see and interact with things more that we agree with. Social media amplifies this in some ways that we've just illustrated, I think.

0:19:49 Thomas: Yeah.

0:20:13 Lindsey: Then once you have all this, then we need to talk about some things like the availability heuristic, which is maybe something that you've encountered before, given your background. The availability heuristic is like this shortcut that we use to evaluate the frequency or importance of certain events.

0:20:33 Thomas: I Right.

0:20:33 Lindsey: And specifically with this, we're estimating the frequency of importance of events based on how easy they are to recall, or how easy examples of them are to recall.

0:20:44 Thomas: So like after 9/11 happens, we think everything's gonna be a terrorist attack kind of thing. Yeah.

0:20:48 Lindsey: Precisely, yeah, that's a textbook example, I think, yeah. Is that after that people thought that plane crashes or terrorist attacks, right, were more common or more of a threat than they actually were. Yeah. And another example that occurred to me was in the lead up to the 2016 election, right? I still have this very vivid memory of sitting there that night and, you know, watching the moment where it was obvious that Trump was going to be elected. And, you Know, the main thought that was dominating me at that moment was like, how the fuck is this possible? Like, I didn't see anybody posting about Trump. I didn't see any ads for Trump. I didn't see any of that. How is it possible that he could possibly be winning this election right now?

0:21:25 Thomas: Yeah. Yeah.

0:21:34 Lindsey: And that's at least in part because of things like the availability heuristic combined with the way that social media algorithms reinforce confirmation bias type things, right? I've got this bubble that's like, partly self-created, and that's leading me to only see posts and things from people that agree with me and ads for things that I already like or agree with. And so my perception of how much support there is for my candidate versus the other one is like vastly distorted.

0:21:44 Thomas: Gosh. Yeah. Yeah.

0:22:04 Lindsey: And this is even given the fact that I'm aware of all this stuff, you know?

0:22:08 Thomas: Yeah. It's still it still happens. And you still you also see it on the other side. Trump supporters love to say, like, I see nothing but Trump yard signs. He's going to win in a landslide like all the time. Yeah, yeah,

0:22:17 Lindsey: Oh my God. Yes. Did you see, did you, there was a, there was like an opinion piece in Fox News that was published recently that the guy was, was like with, with a straight face was making that claim. Like he was making a prediction about the election based on the fact that he saw more Trump yard signs.

0:22:34 Lindsey: It's like, no, It's the anti-scientific approach, that's yeah, okay, and yeah, so things like availability heuristic lead into things like the false consensus effect, and so this refers to the fact that we tend to overestimate in general how many people agree with us. This is true before social media, but all of these things sort of amplify it, right? So once we have all of this, you know, sort of basic effects being amplified on social media, then we're in real trouble. Because now we've got this very distorted set of beliefs about what the social landscape looks like. And those beliefs are very likely to influence our behavior. So I've got three studies to tell you about.

0:22:29 Thomas: Yeah. Perhaps not the most scientific way of doing it.

0:23:18 Thomas: Love studies. This is my favorite part of you coming on the show and

0:23:23 Lindsey: Yeah, yeah, no, so they're really interesting. I think that I think this will be compelling to people. Yeah. So I think that these studies can get us to just the, you know, the social algorithms in these platforms to the downfall of democracy. And just, you you know, three easy studies.

0:23:39 Thomas: Just and just three easy studies. But wait, act now. Yeah

0:23:41 Lindsey: Yeah, yeah, yeah. But wait, there's a lot more studies, but there's not enough time. Right. So the first two, the first few that we're going to talk about, these were actually conducted by at least part of the research team was actually working at Facebook at the time. So Facebook, maybe not surprisingly, employs a pretty impressive number of social psychologists. So, you know people with my training, which the first one to talk about here, this got a lot of attention at the time that it was published from people in my area. It was very controversial, in my opinion, for partly the wrong reasons. But this was a study, a massive experimental study on Facebook that examined something called social contagion, or excuse me, it examined emotional contagion. And so this is a very simple effect. The idea is that, you know, if your friend is in a bad mood and you're around your friend, that'll be more likely to put you in a bad mood.

0:23:46 Thomas: Yeah Yeah. Okay.

0:24:37 Lindsey: So mood is catching.

0:24:38 Thomas: Interesting. Yeah.

0:24:39 Lindsey: And so this was an established effect in the real world. And there was some correlational evidence before this study, the current study that I'm going to talk about came out, indicating that maybe social contagion actually worked online too, right? And this may seem kind kind of obvious to us now. But at the time, like, when studies of social media were still fairly new, it was still a pretty reasonable question as to how influential social media was in comparison to our social interactions in the real world, right?

0:25:15 Thomas: Makes sense.

0:25:16 Lindsey: Yeah, it does, right. I mean, I'm, you real friends with very few of the people that are friends with me on Facebook. And so a lot of social social media sites are primarily made up of people that we have very weak ties to. And we know that the weaker the tie we have to someone, the less influence they have on us. And so maybe social media is really not that important. And there was also some evidence that things like persuasive appeals to change people's opinion were less effective in online formats than they were in the real world. So it was a real question at the time as to whether these basic effects like emotional contagion would actually work in online social platforms. So this research team said, well, you know, the thing that could resolve this is if we used experimental methods, but actually did it on Facebook.

0:26:05 Lindsey: And so they did a very large experimental study using about 700,000 Facebook users. Yeah, that's not even the biggest study we're going to talk about Um.

0:26:13 Thomas: Wow. That's a lot. Oh yeah.

0:26:19 Lindsey: But yeah, they randomly selected uh, 700,000 Facebook users as study subjects, and they exposed them to one of three experimental conditions. So, so now you're gonna see why this is controversial. So participants were randomly assigned um, to either see a normal version of their news feed, so nothing changed for them.

0:26:39 Thomas: Oh, I stupidly missed this. So this was done by people who are able to pull the levers on these, these things. Okay, gotcha.

0:26:41 Lindsey: Yes. Yeah, so They had people, yeah, they had Facebook employees doing this, and they were able to change the algorithm.

0:26:51 Thomas: Gotcha.

0:26:52 Lindsey: Yeah, so in the control conditions, nothing changed about people's news feeds. But then in the two experimental conditions, they manipulated news feeds in one of two ways. So either they slightly reduced the amount of positive content that they saw on their news feeds, or they slightly reduce the amount amount of negative content that they saw in their newsfeeds.

0:27:15 Thomas: Did they try to do it that way? So it didn't look like they're increasing the negative or positive, was that a way to try to look less sinister? Yeah.

0:27:20 Lindsey: Yes. I think it was a, I really think this was like a good faith attempt to be responsible in the way that they investigated it. Because again, I mean, these are, this is a team of people who have been, you know, trained as social psychologists, presumably they care about ethics, you now, and they want to get really high quality evidence in this platform, but they don't want to do things that are actively harmful. Harmful. And so they tried to use like a just a very, a very subtle manipulation. And as you as you point out, they didn't increase the negativity or anything. They did the opposite, I think, in order to minimize harm, and to be responsible. So yeah, so positivity reduced condition, negativity reduced condition and control. And then what they examined was, was during the week where people's newsfeeds were manipulated, did their own posts change in terms of the emotional words that they were using in their own Facebook updates?

0:27:28 Thomas: Okay.

0:28:15 Lindsey: And they found that if they were exposed to the positivity-reduced newsfeed, that people used slightly fewer positive words and slightly more negative words, right? And that's a validated measure of sort of internal states. In other studies, it's been established that, you know, rate of using positive and negative force is a reflection of mood, essentially. And so, yay, they did it.

0:28:26 Thomas: Yeah.

0:28:39 Lindsey: They experimentally established that you can get emotional contagion in online social networks, right? Now, there were two reasons that this was controversial. The first one was that some people were concerned that maybe the actual thing they did in their study was really unethical and harmful.

0:28:56 Lindsey: But like a 10% fluctuation in how much negative or positive content you're seeing, like, that's just, that's consistent with what you're going to see on social media in your normal life without manipulation. The only difference here is that they're, they're doing it systematically through random assignment, and so you can actually infer causality from it, which, you to me is like, I don't think there was necessarily anything think problematic about the actual thing they did in the study. But the other thing that I think, you know, was worth raising a red flag about was that they didn't get informed consent.

0:29:04 Thomas: Well. Right. Oh.

0:29:28 Lindsey: So nobody knew they were in a study. And the reason they could do that, I bet you know why they

0:29:32 Thomas: Because of the terms and conditions, yeah. Human.

0:29:34 Lindsey: Exactly. Yeah, they have part of the click wrap, click wrap agreement is that you consent to be involved in research by being a Facebook user.

0:29:41 Thomas: Experimentation. Yeah.

0:29:43 Lindsey: Exactly? Yeah, it's true. And so Like, yeah, that's like, that really, I think, drew people's attention to fact that, man, we are agreeing to a lot more than we realized here. The second thing I think that's worth being concerned about, and this is how this is the first step toward the toppling of democracies, is that it really does provide proof of concept that Facebook impacts us, it impacts our psychologies, it impacts our emotional states.

0:29:44 Thomas: Okay. Sure.

0:30:15 Lindsey: And that wasn't necessarily something that we knew for sure at that point, and now we do.

0:30:22 Lindsey: And it also shows that, like, depending on who is tweaking the dials of those algorithms, you know, I mean, this was a real subtle effect done by responsible people, as far as we can tell. But if somebody wanted to do something not responsible and unethical, they have the tool to do it. That's something I think is worth being concerned about. The second study, and again, this overlapped in terms of some of the research team, did something similar. So they manipulated something about Facebook to see if it had an influence on people's behavior. And so what they were looking at in this study was voting behavior. Behavior. So they wanted to see, this was this took place in the lead up to the 2010 midterm elections. And they wanted to see if they could change something about Facebook to make people more likely to vote.

0:31:09 Lindsey: Which sounds nice, right? I mean, that's a, it's good socially responsible thing to do. Get people to vote more. Oh, Thomas, you remember that one time when you said that you were never in a social psych study? Were you on Facebook in 2010?

0:31:16 Thomas: Ah. Yeah, yeah. So.

0:31:26 Lindsey: Yeah. So congratulations. You were in study.

0:31:30 Thomas: This was a 100% Facebook user study, okay.

0:31:33 Lindsey: This was all Facebook users in the United States during 2010 that were 18 years or older. Yeah. Yeah, so all of those people, and that was over 60 million users, were part of the study. And they were randomly assigned to be in one of three groups. So one was a control group, nothing changed for them. And then the other two were different sort of approaches to persuading people to vote. And they're very, very similar. There's only one difference between them. So I'm going to start with the more expansive one first. So in this condition, people were shown a statement at the top of their newsfeed that encouraged them to vote, and it provided a link to their local polling place so they could get information about where and how to vote. So that's nice. And then it provided a clickable "I voted" button that had a little counter next to it that showed how many people had self-reported voting.

0:31:40 Thomas: Yeah.

0:32:33 Thomas: They still have that?

0:32:34 Lindsey: I think so. I was trying, I think they might be doing another study because I'm actually seeing something different, but there's some kind of voting information there. And then this last piece, this was the only thing that was different between the two experimental conditions, they showed some people the faces of up to six friends who had clicked the I voted button. So that was the only difference between the two experimental groups. Everybody saw all that other stuff, but one of them saw up to 6 faces of friends who'd clicked I voted and the other group did not see that.

0:32:34 Thomas: Right, pretty sure they do. Yeah. Right.

0:33:07 Lindsey: And so what they're doing with these manipulations, like, you know, with the informational link, they're removing barriers to voting, right? And that's helpful. If you remove barriers to a behavior, people are more likely to do it. And then the other pieces of this are sort of establishing voting as a social norm. And since we are social creatures and we like to conform to things, you know, invoking social norms is a persuasive way to get people to engage in behaviors. And it's more powerful if that norm is enacted by your social group specifically. So not just the little counter of how many Facebook users voted, but your friends specifically it voted. All right. So those were their three conditions. And then they evaluated the impact of being in one of these three conditions on three outcomes.

0:33:53 Lindsey: So one of them was self-reported voting behavior. So they just measured this by whether people clicked, I voted. One of them was whether people clicks the polling place link, which is a nice proxy for whether they were trying to get information to actually vote. And the last thing, and this blew my mind. I had a mixed reaction to it, but part of it was envy. They looked at actual voting in the election. And they did this because, you know, voting records are publicly accessible information, right? You just have to request it. And so they got voting information about the 2010 election from over 6 million, or they got over 6 billion records from 13 states where it was most. And then they were able to match those voting records to Facebook profiles. And so they were actually able to look at, where they got the most increases in voter turnout, like according to the different experimental conditions.

0:34:22 Thomas: Well.

0:34:53 Thomas: Remembering 2010, we got crushed because Democrats didn't show up to vote, so I'm gonna go ahead and say Facebook caused that somehow.

0:35:02 Lindsey: Yeah, I was thinking about that. Like, so, you know, they weren't targeting people based on like party affiliation or anything. But like, I don't know, or who's more likely to be on Facebook? Is there a?

0:35:07 Thomas: Yeah. Yeah, in 2010, I'm sure there was an ideological skew towards younger people, or whatever, demographic skew towards younger people, which then would overlap a little bit with being more liberal, but who knows? I'm sure this wasn't the reason we lost, but What did they find? What were the actual results?

0:35:19 Lindsey: Yeah. Mm hmm. Yeah, okay. Okay. So yeah, so people were more likely to self-report voting by about 2% if they were in the condition where they saw their friends' faces in addition to all the other stuff. So we got self-reported behavior. They were about a quarter of a percent more likely to click the polling place link if they were in that condition where they saw the friends' faces compared to the other two. And then this is the main event. They were about 0.4% more likely to vote in that condition versus the control group or the other experimental group. And this was super interesting to me. So again, like that second experimental group, they saw everything except their friends' faces. They were getting the social norm through the counter. They were getting the information for the polling place. Voting behavior was identical in that group versus the control group.

0:36:19 Thomas: So, it's just seeing the faces that's doing The heavy lifting here. Hmm.

0:36:22 Lindsey: Yes. Yeah. Which I was surprised at. I figured they would get, you know, a little bit more with each sort of part of the manipulation, but that did not seem to be the case. I know that 0.4% doesn't sound like that much. But again, there's over 60 million people in The study. And so their estimates in terms of how many votes they produced as a function of doing this experiment was like on the order of hundreds of thousands of votes, if not a million. An impact.

0:36:25 Thomas: Yeah. Yeah. Yeah.

0:36:54 Lindsey: That is a real thing that they did.

0:36:53 Thomas: That Is An impact, uh, So what are they currently doing with that? I see a lot of stuff, do we know what they're currently doing? I see a lot of, "register to vote," I even though I'm registered and vote. I've, everywhere I turn, and I can't be too mad at it because I really want everybody to vote, but it is getting a little tiresome. Every, every single, everything I, I go to, every social media, everything is like, "Are you registered to vote?" Like, yes, yes I am. Can I do one toggle for this? Like, yes, I already did it. Turn off all the stuff.

0:36:56 Lindsey: Yeah. Yeah, yeah, yeah.

0:37:26 Lindsey: I had that thought the other day too. I really wish that they gave you the option to self-report so they could stop showing it to me. That'd be great.

0:37:31 Thomas: It would be nice. Yeah.

0:37:32 Lindsey: Yep. Yep, I did. I did vote.

0:37:34 Thomas: Yeah, but it's okay. I do appreciate that they're trying to use this superpower to maybe make more people vote. So that's good.

0:37:43 Lindsey: I appreciate that too, because it looks like it matters. Oh, and just, there was one thing I forgot that I think is important to say. They did another socially responsible thing here, which is that they put the vast majority of their participants into the condition that was most likely to have the biggest effect. So it was something like 98% of the participants were actually in that message or that group where they saw their friends' faces, which you can do if you have 60 million participants, right? Like they –

0:37:46 Thomas: Yeah. Oh, okay. Yeah. Yeah, because you still have enough, a statistically significant sample in the other groups. Yeah, I

0:38:17 Lindsey: Exactly. Yeah, they still had over half a million people in the other two groups. –

0:38:24 Thomas: Hear the envy in your voice. This is this must be getting as a psychologist or anyone who wants to do these experiments. This must be a goldmine.

0:38:27 Lindsey: Oh, my God, I know. No. Yeah. Yeah, that first study we talked about, the emotional contagion, I saw that presented at a social psych conference, I think, before the paper was out. And was like, man, I need to work at Facebook. These people got data. It's awesome. All right. So conclusion here, right? They demonstrated that Facebook matters for people's behavior. And they didn't do this in a lab. They didn't do this with some sort of bullshit behavior. They did it with something that really mattered. They they created votes that wouldn't have happened in an actual election, um, and so, to me, right, that that goes a long way toward toward making the point that we need to take this seriously. We need to take the impacts of what's happening on social media seriously.

0:39:17 Thomas: Yeah.

0:39:18 Lindsey: Uh, and again, to go back to that that first point, right, it suggests that you need to, to the extent that we can be comfortable with this power we have, to be comfortable with the people who are sort of in charge of the dials. And exactly.

0:39:26 Thomas: Oh, which I am not.

0:39:30 Lindsey: Zuckerberg's incentives are not aligned with mine. So I've got no trust in that. Okay. So the final study, think, right? We're fairly well along the line of our path to the destruction of democracy. You feel like we're closer

0:39:42 Thomas: Yeah. Well, that last one gave some plus points. You know, I mean, just along those narrow lines that increased democracy. But I suppose that effect being used for ill would definitely potentially reduce democracy quite a bit.

0:39:44 Lindsey: Yeah, Yeah, I mean, what if somebody at Facebook decided, like, you know what, let's remove all of those things for everybody that

0:40:05 Thomas: Take away all their positive words and vote.

0:40:06 Lindsey: Yeah, Yeah, or I just, you know, I mean, if somebody really had a stake in trying to swing an election, and it was along partisan lines, like this suggests, that, I mean, there have, there have been plenty elections that have come down to fewer than a million votes, votes, they could make tangible differences in this if they wanted to specifically target these kinds of manipulations at particular groups.

0:40:30 Thomas: Yeah. Yeah.

0:40:32 Lindsey: And that's disturbing. Okay, so third study, right, this is our last step. Again, this was on Facebook, but the difference between this study and the previous ones is that this was just, you know, a researcher who had a Facebook account and recruited some confederates, some research assistants that also had Facebook accounts. And they wanted to see, like, you know, can we create meaningful change in a local election by, you sort of putting out persuasive appeals on Facebook. And so this piece of it, I think, is important because, Now, the impact of a technique like this doesn't depend on somebody working at Facebook. This is something that anybody could do.

0:40:34 Thomas: Yep. Yeah. Well, does this mean just posting status updates or something? Because that's important because we're somehow already late on time. But this is also different than advertising, which I don't know how well that works. Maybe it works. Maybe you could tell me. But like, that's a different thing than, you know, people posting comments, statuses, all that, which I'm sure will come, will be important later.

0:41:19 Lindsey: Yes, yes, Yes. Yeah, it is. I think that you're right. And actually, yeah, the methods that they use here, I think, specifically speak to that to some extent. Okay, so the methods here, I think are kind of important, and they're a little bit convoluted. So if we are short on time, then that might

0:41:58 Thomas: You can do it. Explain quickly to us. We'll get it. We're smart. Have faith in us.

0:41:59 Lindsey: Okay. All Right, um, so, uh, researcher came seven confederates, right? So we've got seven people that have been randomly assigned to post Facebook statuses in the lead up to an election in Texas, right? And they were randomly assigned to use like different persuasive strategies. And the critical thing about what they did here is that they were tagging specific friends in the status updates. And so people were sort of being singled out, go vote, essentially, right?

0:42:34 Thomas: Okay. Was this one just going and voting or going and voting for a specific like person or thing or

0:42:40 Lindsey: Just going to vote. Yeah, just to get out the vote type thing. So this was, yeah, this was in the lead up to 2014 election. This happened in a particular county in Texas. And so each of these Confederates was randomly assigned to use one of four persuasive appeals. So they had a control group. Andso these people didn't use use any messages at all. They had a civic duty condition. And so in this condition, the Confederates were posting, voting is your civic duty. Just a reminder to friends A, B, and C, the early voting ends on this date. And here's a link to the polling place. They said, voting records are public. Thanks to friends, A, B, and C for voting in past elections.

0:42:40 Thomas: Oh, Okay. Oh, Wow.

0:43:22 Lindsey: And then they gave the information And then they had a shame condition. Isn't that fun? So they said voting records are public. Records show that friends A, B, and C have not yet voted this year. I know. And this was true, by the way.

0:43:37 Lindsey: Because again, voting records are public. And in this county, I don't know if they do this everywhere. But they during early voting, they actually posted updated records At the end of every day during early voting, and so they, the the researchers that were involved in this were like updating their information about their participants constantly throughout the study. So it's, there's so many things that I admire about like the work that went into this, but, um, so people are tagging their friends and these things, and the the long and short of it is that this had fairly profound impacts on voting behavior, right? So they, their outcome was, you know, what turnout out did they get in these different conditions of seeing these different persuasive appeals. And the strongest effect they got was in the shame condition, actually.

0:43:45 Thomas: So Wow. Okay.

0:44:20 Thomas: For

0:44:20 Lindsey: And so 76% of the people that were exposed to those shame appeals voted. The next most dramatic result was for the pride condition. So that was something like 67% of people in that condition voted. And then the control condition and the civic duty position were both around 50%, which is still a little bit higher than average, I think, for midterm elections, but maybe it's not in this county.

0:44:45 Thomas: The sake Of Time I'm trying to resist making the vote shaming joke, but it sounds like, because it's not voting for specific person, so it could be a little different, but anyway, turns Out shaming people into at least voting instead of not voting seems like It works apparently. Wow.

0:44:53 Lindsey: Yeah. Yes. Yeah. Yeah. So this is pretty dramatic. I mean, think, if everybody in the country had been part of that particular condition in 2000.

0:45:08 Thomas: Would have had like 150% turnout.

0:45:11 Lindsey: It's true, right? Right? Like that's, imagine if we got 76% turnout in midterm elections routinely, That would be insane. The reason that I thought this was a nice last step is that it demonstrates that all you need is a Facebook account to be making pretty profound impacts on real world behavior, like participation in, you know, voting in elections and things. And that's, I think, how, how we get to the kinds of problematic things that were reported in the Zhang memo, right?

0:45:12 Thomas: Yeah. Oh yeah.

0:45:43 Thomas: Yeah, this was, we brought this up last time, the memo of the person who left Facebook and released their kind of their memo, or BuzzFeed got a hold of it to refresh everyone's memory. Yep.

0:45:52 Lindsey: Like, yes. Yeah, exactly. Right. And so there's a couple of examples here that I think line up fairly nicely with a number of these claims that we've been making. So, I don't know. So there's not a lot of overlap in several of these. Like a lot, lot of things that you see are people creating fake profiles to, like, the pages of political candidates and, you know, post comments that are supportive of political candidates, or maybe not supportive of other ones. And behavior like this, right, is likely to have pretty substantial impacts on how people are perceiving their social worlds, essentially, right?

0:46:31 Lindsey: How much they're perceiving consensus around particular candidates, and that's super powerful in shaping people's behavior and motivating or shaping people's attitudes and motivating their behavior in these kinds of circumstances. Yes.

0:46:30 Thomas: Yeah so maybe to just drive the point home plainly here, like you've brought up all the examples you use in these studies were kind of like good things, or at least the effects were on purpose by the experimenters. They didn't want to be doing awful things to people. They were kind of trying to achieve what is usually, if you're not a Republican, looked at as a good thing, which is voting and democratic participation and stuff like that. But the point is, and to tie it together with the cognitive biases you outlined before, when these are used for ill and not Not to mention bolstered by tons of fake accounts, giving you an inaccurate picture of how many people believe what and all that kind of thing.

It could be really bad. And it's one thing in the US, you know, it's also, it's another thing where I think, as you said earlier, there there are almost no regulations on any of this. And there's even greater ability for people to manipulate this. So some of the things in the Zhang memo talked about just going off this BuzzFeed write up of it. It says right here, it took Facebook's leaders nine months to act on a coordinated campaign that used thousands of inauthentic assets to boost President Juan Orlando Hernandez of Honduras on a massive scale to mislead the Honduran people. So there's that. There was Azerbaijan. Azerbaijan, Zhang discovered the ruling political party utilized thousands of inauthentic assets, I assume that just means fake accounts, to harass

0:47:10 Lindsey: Yes. Yep. Right.

0:48:14 Thomas: The opposition, etc., etc. There's a number of these examples. So, not good stuff.

0:48:19 Lindsey: Yeah. And, and the point is that that works, right? That you know when people see, like, a bunch of other people sort of mobile, it, mobilizing around a candidate, like that works, that's persuasive. Yeah.

0:48:23 Thomas: Yeah. Well, I'm curious. Do we know for sure? Now, I imagine it is, and I'm not saying it's not, but I wonder how much fake accounts tend to work. Because if, If like when we talked about the get out the vote thing, it was important to see the faces of your friends, you know, that you know at least in some respect. And maybe that would work if you have some fake faces and fake profiles. But I don't know. I wonder, you Know, those experiments, I'm sure, especially in 2010, I would imagine a pretty decently high percentage of people you were friends with Facebook on Facebook with in 2010. You probably actually had seen their face in real life before.

0:49:09 Lindsey: That's true, right, but you only really need the, the fake accounts to get the momentum started, right? If you can, if you can, like, you know, if people are scrolling groups and they encounter a group that sounds like, okay, that maybe is something that I halfway agree with, and there's already, you know, 10,000 people in there, right? It's that, it's that little number that says there's 10,000 people in this group that makes you a little more likely to be like, all right, yeah, I'll join this group, see what's going on, and then at that point you have your real friends in those those groups.

0:49:13 Thomas: Hmm. Right. Yeah. Good point. How hard could it be to stop bots in Facebook? Am I? I know neither of us is really like a tech programmer person. I did some of that work, but nothing involving this. How hard could it be to actually get rid of bot accounts? I guess it's, it's a scale issue. I mean, there's a lot of people in the world and all that, but couldn't there be more?

0:49:41 Lindsey: And it's probably a motivation issue, right?

0:49:59 Thomas: Yeah. Could be a motivation issue. A lot of times it's like these bot farms, I don't know how much they're doing this early on, but, um, back in the day when I've looked into Facebook advertising for any, any number of things, my first thought, my First worry, like I tried it for a few things, Tried it for a few, you know, like podcasts back in the day and you get these clicks and you have no way of knowing if these are real. And so there was a weird incentive where Facebook maybe wouldn't mind having, you know, hordes of fake accounts or bot accounts or whatever you want to call them or people with what are they called? Sock puppet account and people like multiple accounts.

0:50:01 Lindsey: Yeah. Yeah.

0:50:38 Thomas: You know, there are these farms that do this who would then click on the advertising to to make it look like you've gotten some value out of the money you paid Facebook for an ad, when in reality, that's not at all any value. And so maybe there wouldn't be a big incentive for Facebook to get rid of that.

0:50:47 Lindsey: No. No, I, and I think, I think honestly that's, that's one of the big scandals of the Zhang memo, right, is that it's uncovering where their priorities are in terms of clamping down on those kinds of issues. Um, so it, it's, I can, I can I plug an episode of a, of a podcast that I was listening to about this? So, uh, Robert Burt Evans' Behind the Bastards has done several episodes about Mark Zuckerberg, but he recently, around the time that The Social Dilemma was released, he released a two-parter on essentially just all the things that were in the Zhang memo.

0:51:10 Thomas: Sure. Wow.

0:51:28 Lindsey: And it was incredible, and the take-home from it was Zuckerberg only cares about cracking down on these things to the extent that it's giving him bad PR. And it's not giving him bad PR when it happens in these countries that like Americans don't know exist.

0:51:43 Thomas: Yeah, exactly. So I just read an article that I think was along those same lines. And it was about how, well, and how easy it's been for for people on the right to cry foul and make it seem like there's some anti-right bias. And so therefore, Facebook almost, you know, gives carte blanche to these very misleading sources on the right, like whether it's Breitbart or any of the, you know, the stupid Ben Shapiro one, all those things, they get a ton of leeway in terms of the just fake crap that they're spewing. And this was also about their response to Trump when he made that tweet, or that statement on social media, that was just wrong about voting by mail and all that, and the, how long it took them to do such a simple, not even really effective counter to that at first, and they've gotten to the point where they're trying a little more, but they had to be dragged kicking and screaming to any of this stuff, so

0:51:44 Lindsey: Yeah. Right. Yeah, yeah, precisely. And that's what's so disturbing.

0:52:50 Thomas: Did we ruin democracy? Is that what you're saying? We killed it. It's over, and

0:52:53 Lindsey: Yeah. Pretty much. I mean, you know, not here, not yet, but in plenty of places. Yeah, I think that I think that Facebook has already contributed massively to the attempts of dictators and people who'd like to commit genocide to mobilize people. And yeah, Facebook is a really powerful tool that allows them to do that fairly easily.

0:53:13 Thomas: I know I asked you this last time, and I imagine the answer is the same since I don't think this was new information to you, but what do you think we should do? This is just big, structural change, yeah, exactly, this is just, we need some oversight. I mean, you don't seem to think we all ought to log off or anything, thing.

0:53:25 Lindsey: Elect Elizabeth Warren. Yeah. Yeah, No, I mean, I think that there need to be appropriate regulations and sanctions. And right now, the only thing that Zuckerberg seems to respond to is the bottom line, right? And so there needs to, to be, um, there need to be regulations in place that sort of realign the incentives with those of having, having functioning democracies. Sure.

0:53:54 Thomas: Yeah. And it's one thing you can focus on like, oh, well, it's going to be really hard for these, you know, octogenarians in the Senate to try to tell them what to do. And like, you could focus on the negative of, you Know, it'S not going to BE easy, I don't think for any for Congress to regulate these tech firms, based on what how we've seen them be terribly ill-informed going into these hearings and stuff. But on the other hand, I don't know that anybody can effectively regulate this stuff perfectly. Obviously, I think that'll be almost impossible. I sure as hell know that it's not working how it is, and that I don't trust Facebook to regulate themselves. Like, that's That's like step one is I don't trust the multi-billion dollar corporation to just, they'll do, like, scouts honor, they'll do the right thing. I don't trust that. So yeah, I, I'm sure Congress getting in there and and writing some regulations and and trying to do some oversight, what could only be better than what's currently happening. Wow.

0:54:16 Lindsey: Oh, Yeah, absolutely.

0:54:57 Lindsey: Couldn't agree more.

0:54:58 Thomas: Okay, so it's not great, but Uh, so the overall, I guess, to try to summarize, you know, the TL;DR of the podcast, I don't know if that's a thing. But, you know, there are reasons that these things are effective, that these interventions, or whatever you want to call them. But Facebook, was this mainly, this was mainly focused on Facebook, wasn't it? We didn't really talk much about the other platforms.

0:55:01 Lindsey: Nope. Yeah.

0:55:10 Lindsey: Yeah.

0:55:25 Lindsey: Yeah, I mean, there have been other studies that use Twitter and things like that, but I, you know, I decided to sort of zone in on this one.

0:55:26 Thomas: OK, sure. Yeah, I just want to make sure I was speaking precisely here that Facebook there's evidence or scientific evidence of things, interventions on Facebook, things they do, measures they could take, or the algorithm could take or people creating fake bot accounts could do, that those will actually affect behavior. And we have evidence that that has happened. And sure, a handful of times, maybe it's good. Maybe you get people to vote, but a bunch of other times it's really bad.

0:55:32 Lindsey: Yeah. Yep. Yes. Yeah. Yeah. And, yeah, I mean, and, uh, to underline one thing, since you made me think of it, is that, um, the effects are, are small, right? I mean, if you, if you talk about, uh, half a percent, right, um, on an individual level, right, you're probably not seeing a perceptible difference in your own behavior. It it only matters because it's happening at this mass scale, which, yeah,

0:55:59 Thomas: Okay. Right. Hmm. Yeah. That's a good point. It's another reason we shouldn't beat ourselves up too much about being on Facebook. I mean, is that kind of what you're getting at? Like if this is a 0.4% effect on the voting one, for example, like how, how, yeah, that's interesting. Okay.

0:56:30 Lindsey: It's, Yeah, We shouldn't beat ourselves up. And also, that was the other part of this that super irritated me about the documentary, is that they were sort of like focusing in on how these people are like spying on you in particular, manipulating you in particular. And it's like, no, the only reason that these things are problems is because they happen on these massive scales and because they mobilize on these passive scales.

0:56:47 Thomas: Yeah. Well, I'm sure this isn't the end of this conversation. I'm, um, but okay. Hopefully that's, uh, more of a picture of some of the downsides. It's true. We haven't even come close to covering all of it, but yeah, some of real downsides of social media, but still, uh, still enjoy the Serious Inquiries Only group, everybody.

0:57:03 Lindsey: Yeah. Yeah. That's the only group that Doesn't Amplify Your Cognitive Biases It's Fine, yeah, yeah.

0:57:13 Thomas: Yeah. And opening arguments of phosphorus and space. Yeah, yeah. Well, we could talk more about that another time. I am curious about more echo chamber type stuff and all that.

0:57:24 Lindsey: Precisely.

0:57:24 Thomas: And the other thing is, you and I, think we mentioned this last time, too. But like during a pandemic, it's got to be the worst time to think like, oh, I'm going to log out of everything. And if you want to go ahead, I'm not telling you shouldn't, but that's particularly tough time to have all of us take it upon ourselves individually to try to fix a problem that we're not going to be able to fix that will only probably have negative effects for our personal lives. Like not being able to interact with people. Yeah, that would be terrible.

0:57:46 Lindsey: I Think That's True Yeah. Yeah.

0:57:53 Thomas: I wouldn't be able to talk to Lindsay about all the science. We wouldn't have been able to rearrange this recording. Wouldn't have happened.

0:57:56 Lindsey: No. The World would be a worse place.

0:58:02 Thomas: Yeah. Yeah. All right. Well, thanks again, as always, for coming on. I know this isn't the last we're hearing of this. I'm sure there will be more.

0:58:02 Lindsey: Yeah. Yeah. Thanks, thanks, this was fun, yeah.

0:58:20 Clip: Thank you so much for listening to Serious Inquiries Only. If you like the show, the absolute best thing you can do is support us at patreon.com/seriouspod. The second best thing you can do, is please share it on social media or by word of mouth. Speaking of social media, follow me on Twitter at Serious Pod. If would like to be Facebook friends, send me a request at facebook.com Thomas podcast. If not up for that level of intimacy, that's fine. Maybe you want to join the discussion at facebook.com/groups/SIO members. Thank you again for listening. We'll see you next time.