 Welcome everyone, thanks everyone for showing up late in the afternoon, after many other things at a time of quite great emotions for everyone on many dimensions. My name's Jochai Benkler. I teach here, I'm co-director of the Berkman Klein Center and a co-author of this book that we are presenting here today. What we'll do is my co-authors, Rob Farris, who's research director at Berkman Klein and has been a colleague and co-author on reports for a decade and on media cloud for five or six years now, and Hal Roberts, who is a fellow at Berkman Klein and has been the tech lead on building this amazing research platform, media cloud on which we built our research, will speak each for a handful of minutes, giving you the general outline. Then we're enormously grateful and fortunate to have two people come and speak with Martha Minow, who is a 300th anniversary professor at Harvard University, which is pretty much the university professor you're making a face at me, Martha, I shouldn't say it. It's pretty much the fanciest thing this university can say of a person. With enormous experience, former dean of the law school, who brings both real experience from not only as a researcher but as a member of the board of foundations that have been working on this, for CBS looking at it on the business side and who's been working very much on understanding the relationship between what we describe here and the overall media ecosystem and the structure of media and news more generally. Claire Wardell, who is a research fellow at the Shortenstein Center at Harvard Kennedy School and executive director of First Draft, one of the leading organizations in this field working to understand what's going on, working to train journalists, how about how to deal with these questions, and both of them have been enormously generous with their time to come and spend time and talk with us about this set of findings and about their own perspectives on what Claire called information disorder and the questions we need to deal with. So let's start from there. Hal, would you start us? The idea of this book was to wrestle really starting just after the election, even though the work started far before that, with what just happened, what the heck just happened in our country, to our media system, to our democracy. How did we end up in this, what we call in the book, epistemic crisis, where nobody knows what truth is, nobody actually knows what fake is, and where we're in a world where 46% of Trump voters believe that Pete say to a survey that they believe that Pete's a gay has some truth to it. And not only does that seem crazy, but we don't even know what that survey means when 46% of voters say that. This leads pretty quickly in post, president posts are during the Trump presidency into this world where 51% of Republicans against say that they believe that the press is the enemy of the people. So the press is one of our main institutions of truth in our civic system. And this is obviously problematic, but also we just don't even know what that means or how we have a civic sphere when our main institutions of the civic sphere don't work for at least a quarter of our population. So what we tried to do initially in the book is to try to work out who the different actors in this world war. What are the things that we can blame on having gotten here? And a lot of these got a lot of attention right after the election. So Facebook, micro targeting, the Cambridge Analytica, Russian hackers, clickbait fake news, all of these online echo chambers, all of these actually got a lot of press. If you look at the data, they got a lot of press right after the election. This is what happened. It's the Russian hackers, they destroyed our system. But then also there's some bigger, longer term, more political and institutional causes that we consider like the right wing media system that's been inculcating since at least the 90s and arguably since the 50s. Mainstream media itself and Trump himself as a sort of a sui generous personality. One of the ways that we consider this is by marketing this on axes of political versus commercial, institutional versus technological. One of the things you notice is that a lot of these sort of fancy explanations like Facebook and Russia are in the technological and commercial corner of this chart. And one of the things that means is they're easy to blame. It's easy to say, it's Zuckerberg, it's his fault or it's the scary Russian hacker. It's harder to say both as a society and as neutral academics that it's actually the right wing media. I don't know why this is blue, but I'm just gonna keep going. All things have been plugged. All right, we're gonna be blue for a while. So how do we answer these questions? So luckily we have a big giant media data analysis platform called Media Cloud. We collect a whole bunch of data. We've been collecting data for almost ten years now. For this particular bit of work we collected about four million stories both from the election period and post-election period. Those are generally online stories about the election. And then we combine that with other data, both data that we collect ourselves from Reddit and 4chan and Wikileaks and other places and from other systems like the Internet Archive, TV Archive. And very importantly we do a combination not only of quantitative analysis where we take all this data and we apply fancy math to it or fancy layout algorithms to try to learn lessons of it. But we use that data and those quantitative results as a roadmap for really deep qualitative dives to understand not just here's a pretty map, but what does that actually mean when you look at the content? Our first and one of our most important findings is about the architecture of the network. So this is a map of the thousand most, there's something down here that I'm kicking. Ah, there we go. This is a map of the thousand most linked media sources during the election where each dot is a media source and they're linked together by hyperlinks between the media sources. So dots that are close to each other tend to link more to each other. They're colored by audience partisanship that we determined by Twitter. So basically the dark blue sites are sites that are tweeted overwhelmingly by people who also retweeted Hillary Clinton. The dark red sites are tweeted overwhelmingly only by people who also retweeted Trump. And then the pink, green and light blue are various shades between. So this picture tells us some really interesting things. First, we see very little cinderite. So there's very little pink in this map. There's just dark red and then green, which is the center. So if you're reading on the right, you're reading far right. There's no sort of middle of the road right here. And also the right is very insulated. So we would expect to see sort of right on one side, left on, or red on one side, blue on the other side. But what we actually see is we see red all by itself. And then we see that the dark blue, which is the far left, pretty tightly integrated with this mainstream media, which is still getting a lot of attention. What we call this is asymmetric polarization. So what this means is the red is basically operating in its own media world, the right. And the left is operating in a media world that's tightly integrated with these mainstream institutions that have a deep commitment to their version of objective journalism, which is certainly problematic, but is also deeply felt and has real constraints on their behavior. On the right, the oldest site on the right there is Fox News, which is 20 years old. And most of them are much newer. And the vast majority of them don't even pretend to obey any sort of forms of objective journals. This is the same view, which is a little bit easier to read. This is just the top 100 sites. And again, you see the same thing. You just see the right really isolated off on its own. And you see the left pretty tightly integrated around this mainstream media, which is still the center of gravity. What we call this is the right versus the rest. We did the same thing for the first year of the Trump presidency. We see the same thing even more emphasized. You see the right moving even further away after the election from the center. We tried the same thing using Twitter data. So this is the same idea. So these are media sources, but now the links are co-shares on Twitter. So we link two sites together if the same user shared the same story from two different sites on the same day. So if two sites are close to each other on this map, then that means generally they tended to be shared on Twitter by the same user. This is really remarkable. This is an entirely different data set. And it looks very similar, but even more dramatic than the previous map. So we're always very skeptical as quantitative social scientists. We're worried that we're just getting some artifact from our own data or from the math. And this is a pretty strong signal that no, there's something actually happening underneath with the people reading the next. What do the x-axis and y-axis mean? Sorry? What do the x-axis and y-axis mean? The rotation means nothing. So the only meaningful, and I'm sorry, this is fast. I'm happy to answer questions after the talk. The only meaningful information in this chart is how close the nodes are to one another. And to some degree, how close they are to the center. So the fact that Politico is right in the center there means that it tends to be the most evenly shared site from all the different sites. We literally, when I generate these, I just manually rotate them so the red is on the right for obvious reasons. So this is the map laid out exactly the same as the previous map. We just size these by Facebook nodes. What you'll see in the previous map, and in this one, is that with the link map, Fox Nodes and Breitbart were about equally prevalent or influential in the right. For Twitter and Facebook, Breitbart gets more influential in Twitter and even more influential in Facebook. It's also true on the left, Huffington Post, for instance, the partisan sites in general become more influential in the social media sites. And finally, this is the same app we just saw. Oh, sorry. So this is the post-election map of 2017 with just the top 100. And again, you see on Twitter, just more and more separation between the sites even after the election happens. So very stark insulation of the right side. Breitbart and Fox remain the centers of gravity. Even though you'll notice also, when we look at Twitter, there are some really far hyper-partisan sites on the right that are quite important in that cluster, including Infowars and Truthfeed. So finally, this last chart that I'm gonna show is in my mind, maybe our most interesting finding. These charts show, let's start with the top one, that shows the number of hyperlinks, for number of hyperlinks leading to each of the sites in each of these partisanship quintiles. So the blue or the left partisanship quintiles, the red is the right. What we see is even in the hyperlinks, we see a pretty normal distribution on the left side going from far left to the other side of the center left. So we see this sort of typical bellish curve. Not exactly, but there's a distribution of how partisan they are even within the left and the center left or how much attention is paid. On the right, we see a very strong skew. So what this says is not only is the right insulated, but within that right, the vast majority of the hyperlinks are going to sites on the far right. So they're going to sites that since our measure is by Twitter, they're going to sites that are almost only tweeted by very conservative users. The center chart is the same thing but by Twitter shares. And the bottom chart is the same thing but Facebook shares. And what we see is even more extreme. We see a little bit of skew on the left as we go down to Facebook, but we see really dramatic skew on the right. So by the time we get to Facebook, almost the only thing people, a conservative site, almost the only sites that are being shared by conservative users on Facebook are the really, really conservative sites. And that's mostly because Fox and Brightport and the other core influential sites are virtually only viewed by the conservatives. Great. So if technology gods are on my side, I can point with this thing. So can you guys hear me? Wonderful. Great to be here. Just to pile on this slide, one more thought on this is that this is what we mean when we say asymmetric polarization. There are clearly two sides but the two sides are not the same. And the right is more insular, it's more extreme, it's more partisan. That's what this says here. This is different from that. And that's not a subjective opinion, that's an empirical observation. And much of what we try to do in this book is to document that and understand what it means and how it's reflected in different behavior. So what we did is we produced a model that helped us think through this and understand different media systems. And this is a simple model based on media public and politicians. And the general idea is that politicians tell the public what they want to hear. Politicians want favorable coverage from the media. The public wants to understand what's going on in the world. They also have a preference for news that is confirming up their own worldview which is self-reinforcing. And this view says reality check dynamic up there, it's cut off on the top. In a world in which media plays a role where they are actively seeking accuracy and objectivity, they put a check on this whole process. Politicians who go far too far in saying things that the public wants to hear are held to account by media. And media doesn't always transfer happy news to the public. Partisan media works differently. It's a fundamentally different animal. Under this circumstance, the various incentives of the players are the same except that media rather than seeking out accountability and seeking accuracy and objectivity is working on producing partisan talking points and narratives. And what this does is this in effect weakens the checks against this information in the system. In some ways it's a happier system. The politicians get to tell more happy stories to the public. The public gets to hear more of what they like. And the relationship between the media and the politicians is a much friendlier one compared to the adversarial relationship we often see in the prior model. These are different. And what this one enables is a very different tenor of narrative and dialogue. We still see conflict within a partisan media ecosystem but the conflict is less about what is true and what is accurate. It's more about what the right narrative should be and we see policing of the narrative within this. This is a time at which Ritebart was going after Fox News during the primaries, trying to pull them farther to the right. It also allows InfoWars to be a part of this system. A key part to this is this is a structural model and it doesn't require or rely upon any personality traits, any differences in character or integrity or media savvy or media literacy or education. It's a model that's based upon the structures that are brought together and Yokai will talk more about that rather than the individual traits. We don't fundamentally believe that conservatives are more gullible than liberals. We believe they occupy different media ecosystems and this by itself explains a lot of what we see. Here's a quick case study to describe what we're talking about. There's, again, this is cut off unfortunately but there's two media narratives that we tracked and documented in the book. They're both salacious and designed to produce disgust and readers and it's about sexual allegations against Trump and the Clinton. The story of Trump was allegations that he had raped a 13 year old. The Clinton is all the different variations of pedophilia stories involving Bill and Hillary. What we see here is the coverage of these stories, the Trump rape stories. On Facebook, you see a demand for that in Facebook. There's an audience that is sharing these stories and happy to do so. Same for the Clinton pedophilia stories. On Facebook, you see a nice appetite for these kinds of salacious stories. As you move to Twitter, the appetite is still high here. Everything else is flat and in the link economy, the more authoritative the view of the media systems based upon the behavior of media sources themselves, you see nothing on the Trump rape story. And for us, the lesson from this is that the underlying demand for highly partisan material is the same on the right and the left to some degree. It's that demand is both there. There's something else at work here and it's within the structures of the media that are the difference. This is how the Trump rape story played out. It started with a story on the Huffington Post that was very, very popular. Other left wing sites jumped into it and what they did is they dug into the facts on the matter and they beat the story down. This is The Guardian, The Daily Beast and Jezebel saying there's not as much substance to this story. This is a red herring. The difference on the right wing media is everyone's singing in unison on the story. This story in Fox News right here, of all the Fox News stories in the entirety of the election, this was the story that was shared most frequently on Fox News. And the story didn't remain in online news. It also ended up on Fox News. It ended up on their flagship 6 p.m. news program. Newt Gingrich weighed in on Hannity as well. And so this, for us, describes the difference between the media ecosystems is less the supply of disinformation, hyperpartisan clickbait, but the mechanisms that either promote these stories and propagate them further or put a check on them. Just a reminder here that Fox News is far and away the most popular source of information amongst conservatives and Trump voters. Way above anyone else, Clinton voters a mixed media diet. No matter how we look at the data and the sources of data that we look for, the ecosystem looks broadly the same. This is based on survey data by the Pew Research, so a completely different methodology than the work that we did and how showed in drawing together in links and Twitter views and Facebook views, but basically the same story. On the left, the popular sources and in the center, CNN, the major networks, you go out to the left, NPR, MSNBC are popular. On the right, it's Fox, it's Hannity, Slimbaugh, it's Glenbeck. These are different worlds. And in the background on this, we've seen several decades of declining trust in media. Back in the early 70s, around 15% of the population said that they had hardly any confidence in media, and that was fairly static across Republicans, Democrats and independents. Over these past several decades, the number amongst Republicans has risen to 60% and Democrats 40%. And there's the background, Yokai. Okay, thanks. So I'm just gonna tease a couple of the additional things. In the middle of the book, there are a lot of rich case studies that are data guided, but give you a sense of how it works. So what I wanna emphasize for purposes of now, happy to talk about things later. We also have a reception later. We'll be around. Any questions that you don't feel are answered here. We'll be around. We're happy to talk. We're gonna hang out. So this is the deep state. We show how the framework of the deep state was very different before November, 2016. It then shifts to being something about Obama holdovers going against Trump and trying to bring him down instead of being about a deep state in Turkey and Egypt and to some extent, bipartisan criticism from libertarians and civil libertarians against the national security state. Critically here, notice the importance of Fox. What happens is for five years, there's a little bit of deep state. Then there's a massive spike when it's adopted by Fox News as a framework. We use a variety of different measures. So up there on the top right is a Google searches, tweets, online links. And what you see is that as long as it's still in Breitbart and a little bit in the old framework of the intercept, it's relatively low. It spikes with the Flynn firing as Fox and Fox business start to use this frame and repeatedly this is 15-second percentage of 15-second segments on TV. Spikes consistently show an interference pattern where Fox and Fox business are coming to the support and then it replaces online and gets replicated. Similarly, Seth Rich, conspiracy of the DNC staffer was responsible for the DNC hacks. At the beginning it looks very much like the story of online crazy sites, some alt-right Twitter handles, et cetera. But look at what happens and how different it is when Fox News jumps in in November of 2017. It completely, I'm sorry, in May of 2015, which happens exactly when Comey gets fired and Mueller gets appointed. Similarly, we see and we have a detailed study on some really cool propaganda work that Breitbart did to force mainstream media to sort of talk about the Clinton Foundation right after the DNC and that's one chapter. But in a later chapter, we show how the same story gets repurposed. Not to be about Clinton Foundation, although it starts with it, but to get this business of Russia got 20% of U.S. nuke during Obama administration and Hannity, thankfully, most often on YouTube as well, but obviously three million a night, goes from Clinton kickback to the real Russia story are these guys who gave Russians 20% of uranium. So you see the repeated pattern across multiple case studies where when it's only online and not on Fox News, it's relatively small. When you see there on the upper left-hand corner, when it's adopted for a month on TV, and again we see here Fox Business and Fox News, it's more. Critically here, the pattern that Rob described comes from years, decades at this point of sustained attack on mainstream media, on professional media in the right wing. Rush Limbaugh in 93 is already leader of the opposition by 96, 37%, 18% are getting their news from Christian broadcasting and from talk radio. We go through a good bit of the media history and describe how changes in technology, changes in law and regulation, and changes in political culture all worked to reinforce each other in a long-term feedback effect to change the fundamental economics of the outrage industry. And the critical thing that distinguishes Rush Limbaugh and then Fox News from right-wing media from the end of World War II until 1988 is that the former was never a commercial success, whereas what happens is Rush Limbaugh proves a business model in the context of ubiquitous media, in this case AM, and then cable with Fox News, that becomes the first mover to shift things over from trying to get a share of the audience by programming to the middle and becoming a little better, to saying there's a large minority here that could be gotten by giving them exactly what they want, which is outrage. Once those people move, the center-left and professional media start to give some interesting reinforcement to the people on the other side because truth becomes partisan. The earliest we have of our images is 2012. We already see a very similar structure because what happens is that by the time Breitbart in 2007 shows up, there's already an appreciable Fox News effect for 11 years. It's almost 20 years since Limbaugh. The competitive pressures to actually gain attention are completely different than those faced by HuffPo, which starts a year before MSNBC even shifts to a partisan model. And it never really takes off on the left when we can talk about why later. We don't want to understate the importance of Breitbart and the Breitbart-Trump interaction. We have a full chapter on immigration and how Breitbart and Trump strengthened each other to make immigration the agenda of the election and frame it in Islamophobic terms. We don't want to understate the critical role of mainstream media. We're still talking only about 25 to 30% of the population who live inside the propaganda loop and only it. This picture of what people associated with Clinton in September of 2016 can only happen with the collaboration of traditional professional media. And what happened with traditional media, this is the prevalence of sentences related to Clinton and Trump, their scandals and their issues. And critically, what you see is that in the teeth of a highly asymmetric propaganda system, neutrality, practiced as neutrality between the parties rather than as objective pursuit of truth through a countable, transparent means of providing evidence ends up being complicit in the asymmetric propaganda. We can talk Russians and why we don't think the Russians flipped the election even though there's a lot of evidence including in our data that they're there and trying hard. We can talk fake news and commercial clickbait and why we think they only strove through the asymmetric system. Our basic point is not that the internet does not provide for a means for fringe groups to mobilize themselves, whether they're people we agree with and we think that it's deeply democratic, for many people here it will be the movement for black lives, whether we think it's abhorrent and something we should push for many people here that will be unite the right. There are affordances of self-organization for fringe groups to overcome the center. But critically what we focus in this book is what transmits those to become population-wide beliefs and here the thing that most surprised us and to us seems to be most contrary to the prevailing narrative of the moment is that it's professional mainstream media both in its professional centrist model and in its highly commercially successful right-wing model is the scaffolding on which everything else is built. Martha, Claire, can you join us and can we hear your thoughts? So first of all, everybody has to read this book. It's incredibly powerful and it's really the first rigorous empirical study that I know of other than the article that you guys wrote that was already published about what happened in the media what happened with the elections. I had a quick sneak peek of the book and I would summarize the contributions as five, some of which you've heard. The first is that it's empirical and rigorous in terms of studying the patterns of media shares and media usage before the election, after second is that I identify this asymmetrical pattern, the different patterns for the right and the left. The second, third is that the argument of the book is that what's happened is not due to technology per se. It's not due to the architecture of the internet per se and the fourth is that there is an interaction with the business models and the economic motivations that has a lot to do with what's developed. And the last one is that the underlying civic cultures matter so there's a comparison that you didn't hear much about today with other countries that are dealing with very similar situations and yet there hasn't been the same kind of toxic feedback loop. So these are all very compelling and persuasive to me but I have a couple questions. So as to, it's not the technology itself, it's hard to disagree with that except for the business model question that surrounds the technology. So when Facebook, just as an example, had its public offering in 2012, it was a change in Facebook that suddenly the internal sources indicate Facebook had to find a way to monetize and it hadn't been doing that. And there was a ramp up to the engagement criterion which is another word for rubber necking. How to get people to be so fascinated outrages another word that they keep being engaged, keep being engaged, keep engaged. And that's just grown and grown and grown since 2012 and that seems to coincide with your period and that's just one of the players. So the, maybe it's not the architecture but the business model connected to the architecture which then affects not just the escalation of the shares but also the news feeds and also then the advertising. I mean it's all intertwined. So that's a question that I asked. Maybe it's not the technology itself but what about the interaction between the technology and the business models? And we know that the biggest employers right now of behavioral scientists are the tech companies that are programming to make sure that we can't let these machines out of our hands. Another question that I have is on the issue of business model discussion, the fight for attention has a dimension here with the explosion of many, many different devices and media and so the transition from mainstream trying to get the middle to narrow casting. I'd like to understand more about that because some people said that that happened with cable to begin with. I'd like to understand how that interacts with internet and social media particularly. And on the topic of the underlying civic cultures matter again I don't know if it's causation or correlation but the same period of time that's tracked here particularly about the rising distrust of media there's very similar patterns about the declining faith and democracy in America. I mean just it's like just the perfect and so I don't know what's cause and effect but there's something else going on here besides simply the focus on the media and maybe they're related and maybe they're not. Just then finally a few random thoughts. I don't know what to make of the polls that say that NPR is the most trusted source. I just it is it regularly comes out that way I disclose I'm on the board of public broadcasting here in Boston but maybe we just kid ourselves but there it is another is that the interactions between Fox and Breitbart and Trump are so important before the election after the election that we know that staffers in the White House try to get a message on Fox so that the president will see it. So to try to understand how that works really as an organ both of the Trump administration but also frankly the internal communication system of the Trump administration since the election. I think that that's at least a question or something to observe and then finally I do think it was very striking in looking at the structural map Rob about the feedback loop the centrality of identity. The centrality of identity. Now political scientists tell us a lot about the centrality of identity in politics in the 20th century and now the 21st but I think that's another factor to explore and that's a change in democratic politics. So those are my reactions. So I'll be brief cause I just want us to have a discussion. I haven't had a chance to read the book yet but I mean there are at least three slides I wanted to come up and cuddle. I mean the slide about mixed methods. I mean if we have to watch another study that's just Twitter data and the fact that it combines social media data with television and other things I just want this to be the baseline for every country in the world to recognize can we start having proper conversations about it. The last two years have been a number of governments having exercises where they ask for evidence so that they can think about regulation. We shouldn't be going anywhere near regulation when we've got such a tiny amount of empirical research on this topic. And you know I'm sure people will read this and there will be flaws like there are with every study but what you've attempted to do which is to actually look at the whole ecosystem and put it in historical perspective. So the slide with the timeline I definitely wanted to cuddle that. I mean how can we really understand how we've got to this space. So I think it's incredible for those reasons and I work at Shorin Scene Center and we're currently we have a lab where we're monitoring a lot of those spaces so we spend a lot of time on 4chan and 8chan and Discord and we see a lot of the tactics and techniques being played out and what we see is disinformation agents in those spaces attempting to manipulate the mainstream media attempting to manipulate journalists and the platforms knowing that journalists are looking at what's trending what hashtag is doing well as story ideas. We see that play out every day in it's not the dark web it's actually spaces that you don't really wanna go because people are anonymous and it's pretty nasty but they're in these spaces talking through what you've shown us in those graphs. And so I think in order for us to understand this complexity and to think about solutions it's exactly what you're saying and to your point to the identity piece. You know for the last two years I just keep saying in rooms like this as academics and journalists we like to think that people have a rational relationship to information it is a hundred percent emotional and it's a hundred percent about identity a hundred percent about worldview social media is about performance we are making performative decisions every time we decide to share something and I think if we don't understand that space we're in trouble and that came out so beautifully as well in what you're trying to say. So I would lastly say is we're also working across the world we're currently working in Brazil on the Brazilian election we see similar issues there around polarization I've just come back from Europe where we like to pride ourselves on public service broadcasting and what that means and how this won't happen in Europe. Every person I talk to is just saying we're three years behind America aren't we? And I think that's not understanding what that looks like but not learning the lessons and the EU parliamentary elections are happening in May and seeing what's happening in Hungary and Poland and seeing similar forces at play and seeing the role of television RT and Sputnik etc etc it becomes no less terrifying I'm so excited this book has come out but it's 2018 we're about four weeks away from the midterms we're not, we haven't really done anything in the last two years we're not really prepared at all for what's about to happen globally as well as in the US and as I say often when I sit on these chairs is how do we take this to a place where we're actually making change happen and Yoko at the end when you were saying you know who I blame is the journalist I was like, oh my God so I spent a lot of time with journalists but going back to that emotion piece so in July we ran a conference at Harvard and we had 70 of the biggest newsrooms in the country and we mapped out their Facebook page numbers on a graph BBC were in the room they have 45 million Facebook likes so they're at the top of the graph it goes down and everybody felt very good about themselves with their huge Facebook followings we then did the same for the top 70 junk news sites which nobody in this room thinks anybody looks at we plotted on the same graph we put them next door to each other it's green red, green red, green red, green red, green red those junk news sites are getting just as much traffic as those sites and this is exactly what you've played out they're two different worlds there is a reason that those junk news sites that don't have to sign up to editorial policy guidelines that don't have corrections policies that have ethics training, that have legal training and they use emotion they use emotion in their headlines and their visuals every single day it is asymmetrical not just in the ecosystem but in terms of the rules of the game and so I agree with you if we had NPR journalists in here and BBC journalists in here I don't know how we really talk about how they would have to shift their work to really play in this space and I'll leave with this story which was about six months ago I was on a panel with somebody from the State Department and we did a project around the French election and they did a project because it was the American State Department and I said so do you take out micro-targeted ads to Russians in St. Petersburg about how great America is and he said, yeah, we pay for dark ads and I was like, isn't that what we complained about but no, but we have pre-roll that says sponsored by the State Department and what I realized at that point is it is asymmetrical on every single level both journalism, the ecosystem and our foreign policy so that's a depressing place to end but everything, yes, yes, yes as either one of you want to pick up any of this rich set of comments there's so much to do I just wanted to pick up on one thing that Martha said about Facebook and media models and I think what we've seen is an experiment in what purely demand side journalism looks like and it's really ugly and when we see good journalism and we see people holding to the truth there's institutions that are intermediaries involved and I'm not sure how we build that into that system so we're very fortunate to have the legacy of the existing media ecosystem and the norms and the standards that they follow still in existence that many, many people pay attention to but I don't know how we hold on to that and that's something that keeps me up at night but even if we, I mean every morning there are 20 news stories about Trump's tweets because the newsrooms think that they still live in an ecosystem where people go to their brand in an algorithmic system and the algorithm will only ever serve you one or two stories about Trump's tweets because they know that that's not gonna work so really we need to tear everything up, start again because by following that we're not covering the absence of the judiciary the fact that the state appointments have been expanded the fact that we're not talking about how the family is making money because the ethics have gone out of the window so but that requires us to completely rethink what we're doing but instead we're taking the same rules which is making money but you can't make money in an algorithm system where duplication does not get rewarded Martha? Well when the legacy media is deciding who gets promoted and who gets a bonus based on how many hits they get everybody's in the same game the legacy media is not playing a different game in just to continue the depressing point people who cover the Supreme Court for legacy media are judged by how quickly they get their story up and then how many hits and I don't know about you but last time I read a Supreme Court opinion it took a while so there's a distortion profound distortion about even the legacy media what I wanted to say though was about Facebook just specifically there's a two part documentary upcoming on frontline on the 29th and 30th of this month I recommend it Cheryl Sandberg when she was hired to be the grown up and create the business model she said we're not a news organization we're not journalists and I know they're having an identity crisis right now because they now have heard enough and they've stunned their own internal metrics to see how much they are a news source and what are they gonna do because they like the money and they like the money and I don't think they're gonna change so let me push back a little bit on both of those claims about Facebook we've had this discussion before and hopefully we'll continue it often and repeatedly we actually were really surprised by how central the large traditional media outlets were by all of these measures and when we combined that's true both in terms of what sites link to which we take as essentially an image of supply the decisions by media producers to link to each other and give authority and credence and get credence and also by Twitter which we see as an image of the supply essentially what people are attending to and so first of all this finding that even on Facebook the most attention is by traditional media pushes back against this a little bit when we focus on the distribution what you see is that a lot of these engagement with the crazy sites on Facebook Clare that you emphasize is high engagement by a relatively narrower base of unique users and work that is gradually coming out that's more empirical about Facebook suggests that it's a particular demographic our work particularly seems to say that in the U.S. again in the U.S. it has a very clear baseline difference in who is doing it and my sense is that Facebook and the survey data that suggests that there is a small minority that sees social media as their main source there is a larger minority that uses social media a lot for news but most people there's a reason why do they say NPR presumably actually because radio still continues to be the thing that most people actually consume they just don't think about it because they're commuting or they're doing whatever they're doing so I'm a little worried that in the concept not a little bit I read our data is suggesting absolutely the internet social media generally Facebook are wrapping themselves around what we're already doing and projecting forward but what's happening today is still very much anchored in mainstream media and I worry that when we overstate now I agree with you about the incentives and I think we need completely different regulation of advertising and how it's used on Facebook because that's the problem of the next decade and maybe in that regard we got lucky but that's the way in which I see our data is pushing back a little bit. I think it's fair statement but I'm just projecting out because it's the advertising data that is now hollowed out for the legacy media and while it's great to see the bump of interest in the Washington Post and the New York Times we have newspapers and legacy media closing down everywhere else in the country and news deserts all over and local news in particular is a catastrophe so it's just projecting out. So what I would add to this is one way that I read our results is we are at the end or at the beginning of a new stage of this progressive revolution that happened 100 years ago where Harvard as it exists now was created over a course of about 40 years journalism, medicine, all of our modern institutions of truth were basically created between 1900 and 1950 and we're highly trusted until sometime around 1990 and then something happened. One way of viewing that movement is in the 1900s the world got really big and we had to start to make sense of it beyond our town to make sense of the world. It's crazy world across many towns and cities. What we've seen across all of these technologies whether it's AM radio or cable TV or Facebook or Twitter is that they encourage, they encourage there are inversions of clickbait but they're all technologies that happen within some political institutional context and the most surprising part of our findings are that on the liberal side we have a sort of successful in some ways indication of how this can work. You can have addicting info and HuffPost and New York Times working together to moderate each other in a relatively healthy way and that's not all of those sites and mechanisms have their own very large, very difficult problems we have to solve but there's some sense of seeking toward their own versions of the truth but what our data clearly show is there's another political context that's happening on the right that's just veering way off and that says to me that the problem if the problem is technology and we can just say let's just figure out the technology let's just go to Facebook and convince them to be responsible that is hard but seems solvable if the problem is 40% or 25% of our country is just fundamentally doesn't feel invested in this progressive world that we've built and doesn't believe in any of these institutions of trust that's a lot harder. So we're going to end not the conversation but this session at 6.30 and go downstairs to have it in a more relaxed form so we have time for maybe one or two comments, questions. To pick up on the question about mixed methods and also to push you a bit more on whether you have a theory about the extent to which online activity influences and informs political views and outcomes compared to offline communication sources and I wanted to clarify when you refer to traditional media or legacy media in your data are you looking at say for instance CNN are you only looking at online content that's CNN or are you also looking at viewing CNN on TV and if so what was your media consumption data for the offline mainstream media analysis? So we only look at TV on the case studies we don't have access to it on the larger scale it's all the larger scale images the big find is more based on other people's surveys and on our detailed observation online we do use the TV archive to do tech search and then guided research for purpose of looking at TV across the major channels for the specific case studies we don't have when you say offline you mean activism on the street or you mean TV? So have we made judgments about impact in the world of online and TV? No I don't think so did we do that anyway? Well I mean we don't have we there's a gap right so we can look at our own at our web data or even Twitter data where we have very good metrics about exactly what people are doing with that content we can look at survey data to get a sense of what people are viewing in their cars or while they're making dinner we are very aware of and always questing but it's very hard to fill that last gap where we try to understand what is the actual impact we've done this in some of our previous work you can do sort of forensic work where you can talk to people so we wrote a paper on network neutrality and we can actually call people and you can talk senators or aides and say hey you seem to make this decision at this point why did you make this decision? But that's a gap that we're aware of and always trying to close but it's really really hard. Thank you for this. My name is Marcy Merninghan I've been around here for a long time and in 2010 the Kennedy School published a report that my co-author Bill Bowie and I wrote about corporate accountability with the rise of social media and it was published by the Center for Business and Government and when we did our research then we were asking people in companies as well as advocacy groups about Facebook and Twitter and virtually nobody knew much about either one so that was only eight years ago and it's really remarkable to think about where we are now in this turbo charge system. Ithiel DeSola Poole wrote a book called Technologies of Freedom and I wonder there are sort of two questions here and what I'm putting forward. One is that some of what is going on right now could be explained by the inevitable lag between what was then and what is now. What was the current reality is very different than our regulatory and economic systems so the point that Martha was making about CEOs and others are judged on the basis of clicks is an old business model that has yet to fully adapt to the new reality just as the regulatory system exists for yesterday's media environment so I wonder what that those charts will look like in a few years as we catch up with that on the lag side. The second point is and this was part of certainly it's certainly my life's work but it's part of what you touched on earlier. Maybe it's also true in terms of the erosion of civic culture that profound power shifts affecting the economy and politics have left a lot of people feeling voiceless and without any agency and that kind of alienation is toxic as we've seen. It's a ripe organizing feagal for terrorism for authoritarianism and all of that and all the emotional appeals to the game is rigged and you're doing the right thing and it's their problem. How do we do that? How do we address or start to reweave the body, the political and civic fabric or heal the body politic in a way that discernment can be more present and what people see locally as they participate in these global systems at least ground them in some sense of truth and agency where their voices are heard. Thanks. Martha, this may actually, no. No. So this also brings back a point that Martha you had put earlier on the table and that I think we never got to which is the question of what's the relationship between politics between the construction of identity and the long term declining trust in institutions and what we're observing today. And part of what we try to do in the chapter on the origins of asymmetry is connect the emergence of a religious identity as a pillar of the Republican Party and how it connects to the rise of Christian broadcasting as a critically identity-paced pillar. The strategy of the Nixon campaign and thereafter of reshuffling the parties around the Southern strategy with white identity politics at its core and the change in the alliance that made the Republican Party and that was part of the reshuffling of North and South around Democrat-Republican had a lot to do with why it is that these media, these changes in the technology and the regulation found a ready audience. And so that's a story that we tell about how identity gets used for politically strategic purposes and then creates the audience onto which then the business model is able to build itself. And I think that's critically important. And the question of distinguishing the role of media as opposed to background culture and the decline of trust in institutions, I suspect you have at least as much, if not more, than us basically on this, based on this, just look at the various dimensions of it. But again, we have this dual effect. Rush Limba has been talking about the four corners of the seat for decades, government, academia, science, and the media. So there's a component of that, but there's also a component on the left of critique of objectivity, critique of authority. What on earth does this mean against the man? All of this, these together play out with just a deep raising of the baseline. So I think what we see when we look at the data on how many people have almost no trust in media is a combination, I think, of this background shift away from accepting authority towards being generally more skeptical about authority. The deviation between the two is likely, at least in part, a component of the realignment of identity politics in America and the feedback effect with media that propagandizes to this community even more. But that's a guess at a very broad level. We should go have fun. Thank you both so much, Martha and Claire, for coming and spending time with us. Thank you. And thank you all for coming and spending time with us. We're gonna hang around and talk as long as you wanna talk within limits.