Nate Persily and Josh Tucker discuss the impact of social media on democracy and share their research. Nate is a professor of law at Stanford University and a co-director at the Stanford Cyber Policy Center. Josh is a professor of Political Science at NYU and a faculty director at the Center for Social Media and Politics.They are the editors of the recent volume Social Media and Democracy: The State of the Field and Prospects for Reform.
Social Media in Politics
Over the past ten years social media has reshaped politics. Fake news and political disinformation have become a part of the political discourse. But social media has also brought about meaningful change through the #metoo and #blacklivesmatter movements.
Social media has allowed dissident voices to express themselves in authoritarian regimes, but it has also given a platform to anti-democratic views in Western Nations. It has reawakened our sense of fairness, while it has brought to light some of our darkest demons. In the final analysis, social media is both a problem and an opportunity. And your outlook probably depends on the last headline you saw on Twitter or Facebook.
Introducing Nate and Josh
Nate Persily and Josh Tucker are at the forefront of conversations on the role of social media in politics and its influence on democracy. Nate is a professor of law at Stanford, but also has a PhD in political science. He has long been an expert in election law, but has also become among the foremost scholars on the politics of social media and the internet. Among his many roles, he is the co-director of the Stanford Cyber Policy Center.
Josh is a professor of political science at NYU. He specializes in post-communist politics and is the Director of NYU’s Jordan Center for Advanced Study of Russia. But he is also a faculty director at the Center for Social Media and Politics.
Together Nate and Josh edited a volume called Social Media and Democracy: The State of the Field and Prospects for Reform. It is available to download on the Cambridge University Press website. I highly encourage policymakers, researchers, and anyone who is curious to take a look. It features important contributions from well-known scholars such as Francis Fukuyama and Pablo Barberá on a wide range of relevant topics.
In this conversation you will learn why Nate and Josh are at the forefront of research on social media. They rattle off multiple studies their teams conducted that produced groundbreaking research. Now, I have read many articles about the ways social media influences politics, but this is my first podcast where I really grapple with the challenges of the internet. I was fortunate to do so with two of the field’s most important researchers today. This is my conversation with Nate Persily and Josh Tucker…
Nate Persily and Josh Tucker, welcome to the Democracy Paradox.
Thanks for having us.
Thanks so much. Pleasure to be here.
Your research is much broader than just social media. Nate in particular is described as among the foremost experts on election law. And I’ve read studies Josh has done too on elections in post-communist nations. How do you see the connection between social media and elections? How do they interrelate?
For me, the entry into social media and democracy was from the standpoint of both campaign finance and political advertising. And so I had for some time been working on the question of how new technology will change the regulation of campaigns. And as we approached, really right after, the 2012 election had been, thinking about how the platforms in particular, were going to become the main regulators of campaigns. And so my work in campaign finance naturally led to start thinking about how Facebook and Google were going to essentially replace the FTC as the main regulators of political communication.
At the same time, I also was working quite a bit on issues of polarization because that comes up in the context of election regulation. But, at the time, when I was moving to Stanford, from Columbia eight years ago, was editing a book on solutions to political polarization in America. And one of the big questions was how the changes in media were affecting that polarization. That book, Solutions to Political Polarization, the way we talked about political polarization then we should be nostalgic for it compared to what we have now. But both of these areas, about regulating campaigns and the sources of political polarization, were for me, the entry point for thinking about social media and democracy.
Part of the fun of working with Nate on this book and working with Nate generally is that we do come at this from sort of very different entry points and very different backgrounds. So my story of how I got interested in social media and its relationship with democracy actually comes from, as you mentioned, Justin, my work on post-communist politics. So I had written a lot on what were called the Colored Revolutions that took place in the 2000s in Central and Eastern Europe, and the former Soviet Union, which was a series of protests that took place after instances of electoral fraud, also hearkening back to a time, when people claim there was electoral fraud in elections, there actually was electoral fraud in elections.
And so these were a series of large scale protests that took place, some that led to significant changes in regime trajectories in Eastern Europe. So, I ended up being on a number of panels at academic conferences talking about protest in post-communist countries. And one of the things that we heard throughout those panels, or that I heard repeatedly, was that this was never going to happen in Russia. That this could happen in Ukraine, it could happen in Georgia, it could happen in Slovakia, but it wasn’t going to happen in Russia. And then one day in 2011, we woke up, and there was something between 200,000 and 250,000 Russians who were out in Moscow protesting electoral fraud. And so I started asking colleagues of mine in Russia, what was going on and what I heard was Facebook.
And this was the first time I began to think about it as a subject of academic study. Because if this was part of what was determining whether or not people were participating in protests, in the technical language of statistics, if social media was an important causal variable on political participation then, if we were going to continue to ignore it, because it looked different to us. It wasn’t something we were used to studying. It wasn’t something we had thought about. If we were going to ignore this, then we were going to have omitted variable bias in our models. And we weren’t going to be able to correctly estimate our models.
So I started trying to learn about what was going on. I started reading more about social media and around the same time, simultaneously, something else happened, which was that I had a brilliant PhD student by the name of Pablo Barberá, who was at the time a second year PhD student in a parties and partisanship PhD seminar that I was teaching. I let the students pick a term paper that they could do on the subject of political parties or partisanship and Pablo gave me this draft of this paper, where he basically said, ‘I think I have a way to estimate the political ideology of Twitter users.’
And what was so intriguing about the paper Pablo had written is that it didn’t depend on text. It depended on networks. It had to do with who people followed on Twitter. And I remember thinking at the time that if he was right and you could estimate people’s political ideology on Twitter this was going to change how we could do social science research, because it was going to allow us a way to estimate meaningful covariates that we cared about on social media users.
And so it had been one thing to have all this sort of data being out there and floating around in the ether. But we didn’t know things about people. And so it was sort of mind opening to me. And it was one of those epiphanies where I just sort of was like, this is going to change how we could do social science research. So my interest in social media as a scholar came from these two vantage points.
And this is still how I think about it today and how I describe a lot of what we do at The Center for Social Media and Politics is that it was one from the vantage point of social media as a variable. In this sense, that was the impact of social media on politics. And in the early days, when we first started studying this, there was a lot of skepticism that social media would have any impact on politics. That’s an argument we don’t necessarily have to make as forcefully now to justify what we’re doing.
But the other was to think about social media as data. And social media as changing the way that those of us who are trying to do basic scientific research about how individuals interact with the political world that social media would become such a valuable source of data for doing this. And it’s not just social media data. It’s sort of an even larger category of digital trace data. It’s sort of fundamentally changed the way that we can actually go about studying human behavior, raising a whole host of new challenges, but also raising just an incredible amount of new opportunities.
One of the things we realized early on when we first started, what was then the SMaPP lab, and has eventually become The Center for Social Media and Politics was that we thought this was just going to be this incredible tool for studying mass behavior. But we really quickly early on began to also learn that elites were going to use social media and it actually became this incredible tool for studying elite level behavior as well. And so I got very interested from those early days of looking at Russia, both in this question of how social media was going to impact people’s interaction with the political world. But also how, at that time particularly, authoritarian regimes were going to respond to online opposition.
Pablo really studied under you as a PhD student.
Yeah. Oh yeah, definitely. Yeah.
That’s awesome. He comes up so often in the literature. He’s amazing.
He’s well-trained. No, he’s great. He’s great. And it really was the juxtaposition of those two things that got me particularly interested in the topic, particularly at that point in time. So my interest in social media and my involvement here is in part due to having brilliant students like Pablo, who suggest important research questions and show ways to innovate in this field.
It also speaks to how PhD students have such an influence on grabbing onto cutting edge research. I had Zizi Papacharissi on the podcast a couple of weeks ago. And when she started studying the internet and social media, nobody was doing it and it was seen as risky for somebody to base their PhD research on that.
I think by the way that it is still, we are still kind of searching about for how studies of social media fit into the various disciplines. So you’ll see, for example, in elite political science departments, you do not have, I think, representation from this field that we should. And part of it is because the field hasn’t quite figured out what are the right methods, what are the right areas where we should be having a lot of study of this. But this is leading to new departments being created, schools of information studies and the like. And this is what happens when new fields are created. The universities need to readjust.
Now to move onto the topic of social media. It’s always been very complicated for me because it seems like there’s a wide discrepancy between the ideal of social media and how it’s actually used in reality. Larry Diamond described it as liberation technology before he described it as postmodern totalitarianism. So does social media, as it is currently constructed today, contribute to a more deliberative democracy or does it fundamentally undermine it?
So we have an article that we wrote in the Journal of Democracy coauthored with Pablo, but also Molly Roberts at UCSD, and Yannis Theocharis. The four of us wrote this piece in the Journal of Democracy, which is exactly on this topic that you’re asking. And in fact, we called it “From Liberation to Turmoil.” And the idea behind the piece was to ask how can social media be both liberation technology, which Larry had called it, famously in his first article about it. And then Nate had written an article in the Journal of Democracy called “Can Democracy Survive the Internet?” And so we asked. was one of them wrong? Like, did we get this wrong? Was it not liberation technology? Was it not “Can Democracy Survive the Internet?
And the argument that we made there was, actually, you can explain how social media can be both liberation technology and ‘can democracy survive the internet’ with two really simple assumptions. So, the first assumption is that social media gives voice to people who lack access to mainstream media. The second assumption we make is that despite the fact that social media democratizes access to information, despite that fact, it is still absolutely a tool that can be used for censorship.
And this builds off of a ton of the work that Molly Roberts has done in her book, Censored, looking at censorship in China. And that when you combine these two, you actually can get the answer to these questions. So let’s think about it for a second. If you think about how this plays out in authoritarian regimes, well, who doesn’t have access to mainstream media in authoritarian regimes? Well, on the one hand, it might be other would be authoritarians. But, clearly, among the people who don’t have access to mainstream media in authoritarian regimes, are democratic forces.
And so that means in authoritarian regimes, social media becomes a tool where would-be democratic activists can find like-minded people, can coordinate, can plan. Like we were talking about a moment ago in the protest after the electoral fraud in Russia and the 2011 Duma elections. This becomes an incredible tool for sort of pro-democratic activists, but like everything else in social media and technological development around media, these are cat and mouse games. So maybe authoritarian leaders, because the nature of authoritarian promotion systems within authoritarian systems, authoritarians tend to surround themselves with people who are not going to be too critical, so maybe there was a period of time where authoritarian rulers were not really aware of this new threat that emerged, but eventually they did become aware of it. And they began to think about how they were going to respond to online opposition.
And then another piece that we published in Comparative Politics that was published with Denis Stukal and Sergey Sanovich who were then two PhD students of mine at NYU. We laid out a menu of options that regimes could have. They could respond offline. Right? And you can think about this as shooting people, arresting people, and you can think about it as dealing with the regulatory environment in which the social media companies act. You could think about it as reacting online, kind of classical censorship, to just remove content like China. And then you can also think about it though, as a kind of new tool of the internet era, which is to react by engaging online and trying to change the nature of the conversation.
And then we see authoritarian regimes like Russia in particular, you get involved with bots, you get involved with trolls, you get involved with these kinds of government armies to try to swarm in and change the nature of the conversation online. So you get these developments of these reactions.
Now let’s think about it in the context of democracies and you port this over to think about, say the United States. Well, you could take a kind of left-wing Marxist viewpoint and say, well, you know, media companies are instruments of the capitalist class and they’re going to quash progressive voices. Okay. So social media gives rise to progressive voices. And you can think of this in terms of Occupy Wall Street, one of the first that got a lot of attention in this regard. But we can also think about this in terms of things like Black Lives Matter, which was a movement kind of born on the internet and has proved to have remarkable staying power. Now, as we see over and over again, and has become a powerful force in the country offline, as well as online.
But if you think about in democracies, other people who are denied access to mainstream media, at least, until maybe fairly recently, we don’t put Nazis on television. Walter Cronkite doesn’t bring on people who are advocating for the overthrow of the U S government. So anti-systemic forces, illiberal forces, anti-democratic forces were by and large, were not having access to mainstream media. So the same thing, the same affordances of social media that help these pro-democracy activists in authoritarian regimes are available to anti-democratic forces in democratic systems. So you may be the only Nazi in your community and in the pre-social media era, you might have harbored Nazi thoughts inside your head, but you don’t act on them Now social media gives you the tools to find other people who are extremists.
And one of the big questions about social media is what does it mean when small portions of the population, small communities, are able to find each other. On the one hand, you might think there’s a tendency that they can then use some of these tools like bots and trolls to amplify, to make themselves look bigger, to make themselves look more like a larger portion of the population than they actually are, which in turn can have a potentially, spiraling dynamic. They look like they’re more important. They’re able to attract more people.
But the other thing that’s scary about social media and extremists is that we’ve learned it doesn’t actually take large numbers of people to cause damage. So we can have a situation where, maybe this is just fringe elements finding each other, but they egg each other on. And then somebody walks into a synagogue in Pittsburgh and shoots a bunch of people. .
But what I would argue is at the end of the day, my answer to the question is that social media itself is neither inherently good or bad for democracy in so far as it’s an arena, it’s a tool in which political actors can contest for power. And because of the particular speed, and this goes to some of Nate’s work about describing particular characteristics of social media, one of them is the speed at which this changes. It really plays into this cat and mouse dynamic. Whereas you may have had decades to figure out how to harness television correctly. In social media, you may end up with these much shorter periods.
Obviously social media has caused challenges for both democracies and authoritarians regimes in different ways. But are these new challenges that social media creates? Or does it simply expose weaknesses that already existed?
So, I think that social media and the internet do present unique stresses on democracy so in the report I wrote for the Kofi Annan Foundation basically on the Internet’s challenge to democracy. I try to identify those things that I think are unique about the technology that posed stresses for democratic governance. Now, that last part, democratic governance, is important. You talked before about deliberation, right? Which is something that we sort of left out here, because that’s the nature of the internet is it’s in some ways less deliberative than maybe face to face contact, or it’s a different kind of deliberation.
And so what is it about the internet that poses unique threats to democracy. I like to say that there’s velocity, virality, and volume. The sheer speed at which information travels, the fact that it’s done through viral peer to peer communication, and then the amount of information that we have in our pockets with our cell phones.
So, the speed at which information travels means that there are democracy relevant threats from this because late breaking falsehoods, right before an election that are not mediated by any elite filter, then could have destabilizing effects. Now, of course, that can also be the virtue because, if there were elite filters that were preventing some of this late news from coming out, then people would be ignorant of it. And we’ve always had October surprises.
But the point here is not just the speed, but the fact that the communications ecosystem now privileges the kind of content which goes viral. And that is a different kind of content than what was privileged in the previous ecosystem. And we know from the social science research on this that it privileges appeals to emotion, particularly outrage. Right? Certainly love, you know, that’s why you have so many cat videos in your newsfeed. But outrage is what sells, or what gets viral transmission.
In addition to those three features, you have the role that anonymity plays online, which is different than in the offline world, or in the pre-internet age. And so look, that again is one of the great virtues and it’s constitutionally protected. Publius wrote the Federalist papers, and if you’re a dissenter in Turkey, you want anonymity. However, , when anonymity is secure, people, engage in certain kinds of speech that they wouldn’t otherwise. It also is what gives us the bot problem where we’re unable to necessarily know whether we’re talking to a human or machine. We’re lurching toward a system in which it’s going to be difficult for us. I mean, if you’re talking about deliberation, deliberation with a bot is a very different thing than deliberation among small groups of people.
In addition there is no doubt that some people live in echo chambers on the internet. And this is the Nazi point that Josh was making, that if you are a Nazi living in San Francisco, it is very hard to find common cause with your neighbors. But now you have ready-made communities around the web that you can jump into. And then those who are kind of QAnon curious are able to fall into these things.
The final two points I would make on this regard, monopoly and sovereignty, that the roles of Facebook and Google are historically unprecedented. And so these two companies, now their role in moderating political communication is greater than any institution since the pre-reformation Catholic church. And so the role that they play is unique in that they are becoming unaccountable lawmaking institutions, which itself is a kind of stress on democracy. And the final point. Is this point about sovereignty, which is that generally democracies try to wall off their campaign ecosystem. That is more difficult in the internet age than it was previously.
So you’ll notice that when I mentioned all these different features, I never talked about hate speech. I never talked about fake news. Because fake news is as old as news. Hate speech is as old as speech. It’s these features of the internet ecosystem that then pose unique challenges for democracy.
I just want to add one point on that, because Nate at the beginning started off by saying, in the relationship with elections, what if something breaks late and we don’t have time to get it corrected before the election. And there’s this problem with virality. We think that corrections are good. But do corrections work? And also we need to learn a lot more about the way in which people who just casually encounter information online process that information. People who aren’t actually going out and seeking for information about topic X, but they see an article about it in their Facebook feed and they click on that.
And, I think, one of the biggest issues we have to figure out right now is the extent to which this kind of permeates the system writ large, because what’s different now, again, going back to this access to mainstream media, to people who didn’t have access to mainstream media, is that we lack the traditional gatekeepers. There are still gatekeepers to putting information up on the internet. But it is way, way lower than it was when it was the producer of the CBS evening news who decided what Walter Cronkite was going to talk about that night or the publisher of the Washington post was going to decide what gets into the Washington post.
This is a different world that we’re in and people are just getting access to information that presents itself as news, but comes from a much, much, much wider range of sources in terms of the quality of that information
And Donald Trump himself is now a gatekeeper to the extent that he can retweet anything he wants. At least he could, until he lost his Twitter account.
But I want to come back to a topic you mentioned, Josh. You talked about the evolution of social media and how authoritarian regimes react to it. How have everyday citizens begun to change how they react to social media, how they react to misinformation or even disinformation? For instance, I think of advertisements on television. I’m sure that they had a tremendous impact early on in radio and TV when they first came out. But as a child, everybody just ignored advertisements completely. I would imagine that social media is having the same type of evolution where people are starting to interpret the misinformation and disinformation differently. Is that happening yet?
So the first thing is in order to answer questions like that is you just need way more research being done on this because you need baselines and you need to be able to have comparison levels. So first of all, I don’t know the answer to that question.
I think for sure, I would agree with you and then I’m going to disagree with a lot of what you said. But I would agree with you in terms of people evolve in how they interact with social media. Just as the very simple level of the original interaction with social media was based around texts, now there’s lots of people who are going to make most of their interaction around video. And video and texts have different affordances. And that is going to inevitably lead to people reacting in different ways. We may even lay down memories about texts and video in different ways.
But where I’m going to push back on you is you would think, okay, 2016, we learned there was fake news out there and we learned that you could get something like pizza gate. And your implication is people are going to interpret that, they’re going to see this as ridiculous. But… NO! By 2020 we had QAnon. It wasn’t just one pedophile. It was an entire cabal of pedophiles. People didn’t get better about it. And the problem is what we had long known, going back to Nate’s part, about like what’s new and what’s not new about this.
We have long known that partisanship causes people to see things through partisan tinted glasses. Like if you’re a Democrat, you’re more likely to think that the Democrats are doing things well. If you’re a Republican, you’re more likely to think that the Democrats are doing it poorly even if you encounter the same piece of information. This is about cognitive dissonance. There’s a long literature on this in political science to the extent that you identify with a party, you want to adjust things to bring them into focus so that they make sense in this regard.
But we’ve been running a study, actually, Nate and I together along with a bunch of great scholars at The Center for Social Media and Politics at NYU. We’ve been running this study where we’ve been sending out all sorts of news stories that have appeared just in the past 24 hours to ordinary citizens to look at and evaluate the veracity, so essentially code it as false or misleading, code it as true, or code it as could not determine. And then we also send them to professional fact checkers so we have a ground truth and so we can see when people are getting it right, or when people are getting it wrong, based on how often they line up with the consensus of the fact checkers.
But the thing that’s really germane to the question you’ve asked here is that the number one covariate we find, the number one thing that covaries with people’s ability to get this right, and they get this wrong. We look at all sorts of usual subjects like age and education and interest in politics. The thing that has by far, the biggest effect is partisan congruency. And what that means is that conservatives are really bad at thinking things that our professional fact checkers have said are false, they’re really bad at incorrectly thinking that that information is true, if it has a pro-conservative slant. And on the contrary, liberals are really bad at it when it has a pro-liberal slant. So this effect dwarfs other things that you would think would matter more.
So I’m not optimistic that actually people have gotten better at this, and they’ve gotten better at discounting it. Which is not to say the first part of your question is not the case that people interact with social media differently. And I think they of course will continue to do so.
So Josh, regarding your study, this morning I read a paper that talked about democratic hypocrisy and they mentioned how Republicans were more likely to have anti-democratic views when their party was in power than Democrats. Now that said both people who are far on the left, have those opinions, people far on the right have those opinions, but Republicans tended to have those opinions more than Democrats. Did you find a difference in partisan lean? Or were liberals just as likely to believe leftist opinions, even if they’re untrue as right-wing people?
We had a slightly less instances of left-leaning things that were coded as being false. And so we had sort of wider errors in this regard. There was a slight difference. There was slightly more, you know, more getting it wrong on the conservative side than on the liberal side. But again, the size of that difference was absolutely dwarfed by this congruency difference. The huge difference was between when it aligned with your political issues and when it didn’t align with your political issues, as opposed to whether you were conservative or you were liberal. Conservatives, you do a little worse in this regard, but we also had less data on liberals.
Let me just had one meta methodological point, Justin, on this, because you asked a very important question, right? Is there something in the nature of conservatism either today or historically that leads them to harbor either authoritarian tendencies on the one hand, or certain features of information processing related to democracy. Now, one of the problems in our field is that it sort of blossomed during the Trump years. Now, of course ,there’s plenty of it beforehand. And so we’re living in a particular moment in the United States where most of this research is being done. And so, if the partisan congruency point that Josh mentions and that we argue in the piece is true, then you’ve got a unique universe of the current information ecosystem, which is then going to be driving people’s assumptions about conservatives and liberals.
And so what we need is to have more cross national studies on this because, there is no question, that for the last four years that a greater share of the information that’s coming out of conservative sources is going to have a higher amount of disinformation. And however we’re going to measure that now, that’s because Trump was the president and because his Twitter feed was so dominant and because it echoed throughout the ecosystem. But that’s not always going to be true. And it’s not true universally around the world. And so, for example, in Britain as I understand it, there’s a much healthier share of disinformation on the left.
And so we need to really, , expand the lens to look at how different people in different countries are processing this information. But I would still stand by the basic motivated reasoning and information filtering point that Josh was mentioning about how people are processing this information to make it consistent with their preexisting ideological attachments.
I mean, just to reiterate this. So we had another study out of our lab that looked at who had shared fake news on Facebook in the 2016 elections, which was co-authored with Andy Guess, who had been a postdoc in our lab and who wrote a chapter on this for our Social Media and Democracy book. And we had very, very unique data in that case because this was pre-Cambridge Analytica. And we allowed people who were participating in the survey to share their Facebook data with us if they wanted to. And there was a way to easily do that at that point in time, where they would log onto their Facebook account, so we were able to collect the data.
And so about half of our sample, or half of our sample who had Facebook accounts, agreed to share their Facebook data with us. So after 2016, we could look at who had shared fake news online. And we did find that it was overwhelmingly conservatives who had shared fake news online in the 2016 election. But of course, as Nate just mentioned, the vast, vast majority of fake news out there in 2016 was of a pro-Trump variety. And I’ve seen studies that suggest that it was as much as four to one, if not even higher. And even in that study where we looked and we could identify a few articles that were fake that had a liberal slant, those were shared more by liberals.
But it was ,as Nate’s pointing out, it’s this weird period of time where there’s so much more pro-conservative fake news in the ecosystem. And anytime you do a study like this you’re going to find exactly what we found in that particular study, which was that more of the fake news in 2016 shared on Facebook was shared by conservatives. But that’s in part because it was pro-Trump fake news.
The other thing we found in that study though that was phenomenally interesting was that even controlling for partisanship. We found that the biggest predictor of who was sharing links to these fake news websites was age. And in particular, what we found was that people, over 65 on average, shared seven times as many links to these fake news websites as millennials did. But we also found that this relationship was just monotonically increasing. The older you got, the more likely you were to do that.
So that was a big deal because up until that point in time, I think people had mainly been talking about sharing fake news on Facebook as a problem we were going to fix by doing digital literacy courses in high schools and getting all these young people who are on social media educated as to being able to identify what was fake and what was not.
In answer to that paper that you were reading , some of the research I did before I got into social media was about partisan viewing of public opinion. So again, the idea that right now in time Republicans are more likely to express pro authoritarian sentiments than Democrats are, remember you have a leader of the Republican party in Donald Trump who himself has been expressing pro-authoritarian sentiments.
And my colleague Pat Egan at NYU has put together a bunch of stunning graphs that shows how much Republican public opinion changed on bedrock issues of public opinion related to the Republican party when Trump took over and there are just incredible figures about how Republicans who were all pro-free trade – Trump comes – in support among Republicans for pre-trade craters. Republicans who have been huge supporters of the FBI for years, Trump comes in and starts dumping on the FBI, support for the FBI goes down.
So, it is always important to have these baselines. Whenever you’re asking questions of your guests about has this changed some of this stuff, it’s really hard to say if we only have a snapshot at this particular moment in time. And as Nate points out, this has been a particularly weird moment in time in US history that we’ve been living through lately.
Nate, I’ve got a question for you about free speech. And it’s always difficult, because when people talk about free speech, they wrap it up into the First Amendment. But I get the sense that both of you, not just Josh, look at social media from a comparative lens. So I want to ask you more broadly in terms of free speech, is misinformation or even disinformation a form of free speech?
Well, it is. And you know, the devil of course is in the details here, because part of the argument for robust, free speech protections is that you don’t know at the time the speech comes out, whether it’s disinformation, whether it’s a lie, or whether it’s truth. So sort of famously John Stuart Mill argues about how even some falsehoods are going to be protected because you want to make sure that they lose out in the marketplace of ideas and it helps both bolster the truth that it had this kind of competition as well as to decrease falsehoods .
And just speaking as a law professor, the Supreme Court has upheld certain types of falsehoods, lies as being constitutionally protected. A famous case, U.S. vs. Alvarez, the question was whether they were allowed to falsely claim they’d won the congressional medal of honor. And they struck down the law that prevented someone from saying that. So there are now of course there are other examples of it’s not considered protected speech: fraud, defamation, right, all kinds of falsehoods, literally falsely shouting fire in a crowded theater as being the classic example.
There all kinds of the instances where you can, whether we call them speech or not, it’s unprotected speech in those domains. And so, one thing that I want to emphasize whenever I have a debate about disinformation is that the problem that we throw into the category of disinformation when we think about online harms on the internet, the lion’s share of what people think about when they’re talking about disinformation really is not going to be addressed by drawing a line between what’s true and what’s false. The problem of disinformation is not just some lies that are being propagated online. It’s false beliefs that get instilled in people either through true statements or unfalsifiable statements or the like.
So I often give this example, consider three headlines. One that says, ‘Obama was not born in the United States.’ The second, ‘Trump claims Obama was not born in the United States.’ The third, ‘Liberal pundits disagree with Trump’s characterization that Obama was not born in the United States.’ Now, one of those statements is false, two of them are true. They all have the same effect on the reader And so it’s not as if you have a kind of truth serum for the internet, that it would actually solve the disinformation problem. And the platforms are realizing this. If you look, for example, at the kinds of things that they’re putting labels on, a lot of it is not that it’s false. It’s that it’s misleading or it’s sort of appealing to emotional attachments to what are false narratives and the like. And so this is why it becomes from a policy perspective, an extremely difficult problem.
Josh, as a post-Soviet scholar, I’m sure that you’re familiar with the influence of Russia into the United States, into Ukraine, Estonia, lots of different situations. Do foreign nations have a right to be able to communicate their opinions in other nations’ elections. And if so, is it simply necessary for them to disclose that they’re expressing these opinions? Does that make it okay, or is it wrong for them to get involved at all?
So I want to say a few quick responses to that. One is as an American citizen, I would always, always want my government to prevent foreign entities from interfering in the electoral integrity of our elections. I want my government to protect domestic entities from interfering in the integrity of our elections.
As a social scientist though, I find myself wanting to answer the question of can we measure: A. The ways in which governments are trying to interfere in other countries elections, because that’s important for our understanding of what’s going on and it’s important for public policy responses. But B. as social scientists, when we do basic science we want to see if these things actually have an impact.
And so I think we need to think very carefully when we talk about this broad category of governments interfering in other countries elections. There are as you’re talking about statements. Do they have the right to make statements? Do they have a right to run advertisements? Which is a legal question. And then there are all sorts of more nefarious ways of interfering in elections. And I think as social scientists, part of our job is to try to figure out what the impact is of these different sorts of nefarious ways of interfering in elections.
Well, there’s interesting case law on this, for those who are interested. If you go to election report at stanford.edu, we published a series of recommendations following the 2016 election. And one of them deals with foreign broadcasters and particularly how you deal with RT and Sputnik. And can you distinguish between them and the BBC? How do you deal with foreign intelligence agents and the like? I mean, there’s a lot to say here.
The short answer is that the courts in the U.S. have upheld applying campaign finance bans. So campaign activity and spending for foreign nationals is banned in the United States. Whereas as we know from Citizens United, not just Americans, but corporations have First Amendment rights to spend as much money as they wish on candidate elections. And so this is a much trickier problem, and your question suggests that it is, than people realize.
You can’t, for example, say that no foreign entity shall have access to U.S. markets that might potentially influence the outcome of an election, because BBC reports that put a candidate in a bad light could potentially do that. How do you distinguish between BBC and Sputnik, let alone the Russian intelligence services? And as you’re saying, like, how do you distinguish between even the nefarious sides, of propagandistic sides of things, and Voice of America. And how the U.S. has historically broadcast around the world. In the Kofi Annan report, we argue for a kind of Geneva Convention on this that we need to have good rules that are agreed to from everyone on what is the proper role of one country in another country’s elections.
Now, Josh already mentioned an article that you wrote, Nate, back, I think it was, in 2017, “Can Democracy Survive the Internet?” So, as we look to conclude, I think it’s appropriate to ask, can democracy survive the internet?
The answer is yes, but it needs to adapt. And democracy let alone other governance systems has always adapted to technology. Technology and technological change poses stresses on all forms of governments and democracy is no exception to that. And so, one thing that I think we are really grasping about for right now is how should democracies deal with these multinational monopolistic platforms like Facebook and Google. And no one has come up with a good answer. But there is experimentation afoot. So for example, the Facebook oversight board, which has now started issuing decisions, and will reconsider the issue of de-platforming president Trump is one example of an experiment in this area. I don’t think it’s the answer, but we are going to have to lurch forward in thinking about how the regulators, the information environment, fit into a larger theory of democracy.
Yes, I think democracy can survive the internet. Absolutely. And, I think, you can also ask, ‘Can an autocracy survived the internet?’ These are cat and mouse games, where you have political actors that have long struggled for power and they use what tools are available for them in those struggles. And the same way that television advertising, radio advertising, these things changed things before, technology is changing it. But the thing that’s really different here is, A, the speed at which these things are changing. I think we are in this kind of funnel and that itself is dangerous, both for authoritarian regimes and for democratic regimes, because it’s harder to react to changes.
Look at Russia in the last couple of days. Russia has been rocked by a YouTube video. Now there’s a lot of stuff behind that YouTube video, but nevertheless, a YouTube video. So everything Russia has done, all the things we’ve talked about. The last decade, Russia has come up with ways to deal with online opposition. Now they have more protesters on the street than they’ve had in a decade. Again, because of Navalny’s return and because of a very clever video that Navalny released on YouTube.
So these things change quickly for open and free societies that depend on regulation to mitigate against the potential to exploit the openness and the freedom in their societies. This poses particular sets of challenges, because regulation is a slow prodding process. And you really always want to be careful that regulation doesn’t have unintended consequences that are worse.
And so the process by which these things change, it may be the case that by the time you come up with a satisfactory regulation, the underlying reality has changed so much that the inadvertent effects of that regulation have gotten worse and the regulation is going to cause more damage than it would before.
So I think the thing I would close with, and this is the way actually that we close our book, Nate and I have this concluding chapter where we talk about that one of the things that is just so unique about this era, going back, even at the beginning of what we talked about, at the beginning of the podcast, that my interest in social media came about in part, because I looked at this data and said, ‘this is going to change how we do social science research.’
Well, it’s changed how we do social science research in another way too, which is that social scientists used to do research primarily with data that was in the public domain produced by governments like election results, unemployment figures, or we used to do research with data we collected ourselves. We put people in a lab or we’d run a survey and do these kinds of things.
Now we’re in a position where we refer to it as the best of times and the worst of times for this kind of research, because on the one hand, there’s more data than we ever could have dreamed of having access to about human behavior before. On the other hand, that data, as Nate keeps referencing here, is in the hands of a very small number of private companies that are huge political actors in their own right.
So I would close with a plea for is that these kinds of questions that you are asking us here today, in order to answer those questions, we need rigorous scientific research, the type of research that we talk about in the book, and the only way that this rigorous scientific research is going to take place is if people who are outside of the platforms, who are not employees of Facebook and Twitter or Google and are bound by all sorts of requirements of their contracts to work there in terms of what they can and cannot release publicly.
We need people who are going to put their findings from research into the public domain to inform the necessary kind of policy-making here. We need to think about how that data is going to be made available for public facing research. And that is a role where governments, as they think about reform that is a really important role that governments can play.
But as Nate was saying, this is another place where there’s going to be experimentation. As we begin to think about this, but this is an underlying, issue that’s going to permeate all of these questions we have, which is what is our ability to get good answers to the underlying scientific questions about how things are progressing, how things are functioning. And it’s especially crucial in an era where a study from five years ago may be irrelevant today.
We’ve touched on so many topics and yet we’re barely scratching the surface. It really shows how deep and insightful, social media and democracy, the conversation surrounding it can become. Thank you so much for joining me, Josh and Nate.
Thanks. It’s been a pleasure.