Discussion on technology and how can increase or decrease better turnout, and look at the best way to educate local journalist covering issue, it is an hour and a half. Okay everybody it is time to take your seats please. So our next panel, this is the whole side you will see for the next panel so the theme of this panel is Voter Suppression and voter turnout and in particular the roles that Network Information technologies play in Voter Suppression and voter turnout from a variety of different perspectives so another way to think of it is maybe as peeling back the curtain a bit and getting into the technological underbelly of some of the things that the prior panel discussed. Once again we have a fantastic panel of experts from multiple perspectives backgrounds and disciplines and im not gonna repeat everything thats in the program but just briefly. To my immediate left is kara here is a senior policy analyst in the Economic Security initiative that georgetown law and our center of poverty and inequality she comes from a policy background, we have that jessica who who is a reporter with public a boost cover these extensively. Matt monaco is the Research Director who has studied bots and he is coming to us with computational linguistics background so correcting the other guy was real or the bought. And then finally at the opposite end of the table at least geographically is katie peters who is a cofounder of democracy works shes no longer with it but wall in her tenure there she was instrumental in their acquisition of the bar information project she was named to the forbes 30 under 30 for her works on democratizing access to new voting in voting information we are super delighted to have all of you here and as the Previous Panel did i asked henry one to make a ten to 12minute opening set of remarks and then i will ask some questions and then open the mics to you. So jessica is going to lead us off. Okay, i feel a little underprepared now that you had fancy sides before but i will try to do my best so what, i come at this from a different perspective than the rest of the panelists, i will be decidedly less tacky than the other three but hopefully thats welcome here, not you okay good then will be in good company so, i want to explain first what election land is, some context for the position that im coming at this information from, election land is something we have been doing since 2016 we did in 2016 in 2018 and prepping to do it again in 2020, the premise is that we try to hold data on election data in realtime and disseminate them to hundreds of media organizations so in 2016 in 2018 we had more than 100 local News Partners across the country who are feeding information to, prior to election land the way that the media covered problems especially on election day but even during early voting was they would drive around the polling places and be like is anything wrong and workers would be like no and the drive to the other polling place and now we have information Realtime Data from a couple of networks that do similar systems and they give us realtime information to those calls we can see where the problems are and we can send journalists their in realtime, so the idea if lightest shined at places where there is a problem maybe will be fixed, rather than the next day saying this polling place plead up and we can fix it in realtime, the way that we can do that is by training a lot of journalists on Election Administration i know that some of you in this room have been actively involved in elections so theres things that are problems benefactor not problems machines break it is normal and its not preventing a lot of people from voting theres no point im reporting on it, a poll worker will make a mistake thats normal, lines will be 30 minutes long, thats normal so a lot of what we see our job is cooling down the hysterics because we believe that if we show that there are problems where there are not actual problems and we make them into bigger problems and are people are less likely to come to polls, we are now responsible for over a fleeting confidence theres problems any want to make sure the real before we send local media to cover them thats the role we see far selves and the reason we see ourselves in those participants because we believe and i personally believe that the media is not just an observer of the elections are an active participant and them. So we have to accept that role and work within it and if you think about it the people who are looking for information on their polling place on the position of their candidates looking for information on Voter Registration deadlines are very frequently not looking at their counter registers website because they probably dont even know where it is but they are going to their local News Organization website so its in our interest to make sure that we have local media be reporting the most Accurate Information possible to that and we deeply encourage our partners to do things like go visit the county registrar play with the voting machine so they know how they work so if they meltdown they can authoritatively say what the problem is or maybe there is not a problem or maybe theyre old and machines or miss calibrated so we wanted to cool down the situation so not everyone is in hysterics and we can make sure that we are not unintentionally spreading misinformation through our own reporting so i think that knowing what to cover and what to give attention to has been focused really after now on the candidates complaining that this candidate got more coverage than i did whether you guys are the gay peepers of news and youre giving donald trump more advertising and in that way you participated in democracy thats true but its also true that if we overinflate one problem and dont give attention to another, where we overinflate one problem and dont give attention to all of the ways it can be improved we can be the gatekeepers to things that lead to turn out or dont. So i think you know the trust in the media here is a really important thing i know that the Previous Panel touched on this a little bit for example we are seeing this right now the ap has declined to call the caucus race purity voters not happy despite making a choice and being a pulitzer finalist for doing the exact same thing so when we make calls like this its up to the public to decide how do interpret those things and all we can do is make sure making the most responsible choice. From an administration standpoint we need to do a better job educating journalists on how elections actually work and how they function to make sure they are reporting Accurate Information that is what we try to do so i think just to wrap this all up i think that if the media can take a stronger stance and actively becoming part of the administration everyone will be better served they will have better information before and on election day, i think it will help turn out there are several studies that helped show the more coverage you gave these issues the better turnout it is or the more educated your local populace is i think we should take that very seriously and feel responsible for that information and thats what we are hoping election land will do. Thank you applause nick over to you. Well let me know if you cant hear me. Thank you for having me im really excited to be here. My name is rick manaco, i have been studying misinformation and bots and how theyve been spent spreading misinformation and particularly concentrating an election context so a lot of the work that ive done over the last five years has had a global lens and how bots this, so the topic on the Conference Today is really central to the research im doing and im really excited to do research with all the great people here, im director at the place called the Digital Intelligence lab with a think tank in paul Alto California and what we do is we try to produce public facing research that documents emerging ethical issues at the intersection of society and technology, and what we have done has been document tank state sponsored trolling harassment and disinformation, today and really excited to talk about some of the research that i havent actually been able to talk about before being in the disinformation world especially as a researcher right now i think some of us, there is a problem of disinformation abundant so i thought about so many pieces and studied so many pieces of this mind that i think some of the interesting things so this is a good place to synthesize my thoughts and think about some of the Key Takeaways from some interesting elections and ive studied in the past. Mind so with that mike my intention today is to cover to case studies, one of but increasing voter turnout and one of but decreasing voter porn their turnout. I should be circumspect and say attempting to increase or in attempting to decrease. Because that doesnt think is a big gap in the research right now is i think many of you know. And then i would like to talk about kind of a more passive use case of bots which i think doesnt get a lot of coverage. I think you all are familiar with the idea of a mega funding or amplifying message. They can be used in a more passive way to collect intelligence or collect data online without making noise. So i think thats something worth pointing out. Before we dive into these case studies i would like to just set some common definitions so that we know what we are talking about and were talking about bots i dont think this room probably needs this but just in case if anyone in the audience isnt too familiar. But our computer programs that poses humans online. They are automated programs. And political bots in particular which, we will be talking about today, a buts that are deployed to affect political goal. So social manipulation, opinion manipulation or social persuasion is really the goal in these cases. The builders and the deplores of these spots really want to convince the public or certain portion of the public to believe or act ideally in a certain way. So with that in mind i think the first case i would like to talk about is macedonia or i should say North Macedonia at this point. And basically in september 2018 there was a referendum in the country of what was then macedonia up on whether or not the country should change their name to the republic of North Macedonia. This was the first in a series of steps that composed something called pexa agreement which was an agreement between the macedonia government and the greek government to resolve a name issue. Essentially the greek government had bots. Macedonias acceding to nato or the eu saw for a long time because of the issues that greece has a region called macedonia within its own borders as well. So this is a kind of detente of those tensions. But the first step was to have a referendum locally in the country of macedonia, to consult the citizens and say, is this something we want to do . This was a consulted tory referendum it, was not binding. But it was something that ideally the government was hoping that there would be good turnout for so that they would be acting with the will of the citizens. They were hoping for at least 50 turnout if not a greater amount of turnout. This was a very interesting case study so i did research internally at a Company Called graphico where i used to work and Intelligence Company that specializes in dibss information. So i lead research is internally add graphica on the message on you lectures and collecting facebook and twitter data. There are others such as these transatlantic integrity conveyed condition in the digital corrective lab. Who also covered. What was interesting was we all saw the same thing which was to say they were about ten accounts that were hyperactive, that were producing hundreds if not thousands of tweets per day. That were promoting the idea of boycotting this referendum and were saying look, you shouldnt participate in this democratic process, this referendum isnt necessary, its not valid. You just shouldnt participate. So theres a hashtag, hash dog bikotera, which is macedonia for boy dog. So it is spread by these suspicious accounts about only ten accounts. They were highly automated. My internal analysis determined that as well as the Digital Forensics lab and a great report from the book and reporting network. A bunch of bots and anonymous accounts were producing tens of thousands of tweets in the months leading up to the referendum and encouraging this idea of boycotting. And in any way the referendum went off but the turnout was only 37 , far into the 50 the government was hoping for. Then another interesting thing to note in this case was that a lot of the traffic appeared to be driven by macedonia diaspora who were citizens but residing outside of the country. So boikotera, dot and k as well as some of the accounts that were producing about disinformation and may be true information and not were produced by a bright macedonia living in a way for example. And as a decision as the digital investigation revealed. A quick footnote, i also discovered a nigerian president ial election in the following february in 2019, and there was also a coordinated campaign in that case as well to emphasize again boycotting the election and not showing up. The motivations and the context there were quite difference. There was a but that was producing hundreds of tweets the days, saying dont turn out, dont participate in the president ial elections, nigerians on markers he is not valid. And they came from a place called. Biafra. Its a state that attempted to succeed in the 1970s causing the nigerian civil war but the context was very different. These campaigns showed coordinated behavior that was trying to encourage citizens to not show up for the elections. One case i am really excited to talk about is a case of but that were used to try to increase voters prefer not turn out in the united kingdom. It was actually covered quite heavily in a lot of Media Outlets but still i dont think it gets the attention that it deserves. So essentially how this work was there was a group of activists who were hoping to increase voter turnout, specifically among young people, to vote for the labour party in the 2017 snap parliamentary elections in the united kingdom. But what they decided to do was look, we want to increase voter turnout for a demographic that is all we did to find, 18 yards to 25 years old. This slice of people who we would like to get the turnout. What did these people usually do . Well a lot of them are on tinder, a lot of their minds like swapping out and trying to go and days. Why dont we just brother but thats on tinder and talks to people about their political views . And it seems like we can add to them to vote for the neighbor party, we will do that. If it seems like theyre going to vote for the conservative party, we will just tell them not to come up. This is what they did. What was interesting about this case is the whole time this was a fairly transparent process, the activists and the team of developers that built this bought, they were overt about it. They fund raised on indigo go publicly, and they said this thing was going on and how it worked was if you wanted to, if you agreed with a cause, you could volunteer for a takedown to be connected to this bought and to carry out these operations that i described already. What is it about this is that of course you dont know that you are talking with about when you are. And indeed, this is a case of cyber activity so part of the profile is automated but part of it is human. This means that you can talk with a user if ive volunteer my account and i could text everything i want to text them but one politics comes up the vote will take over and start spreading the messages to vote for the labour party. And then i can take over again. So its pretty insidious and its an interesting case. One thing i would like to note about this case is that it received positive press coverage of fairly unanimously so bbc in the garden and a lot of people wrote about these spots, saying this look at this great thing that these labor activists did and indeed, i should note as well that that demographic of people that they were targeting, 18 to 25 years olds, showed the high turnout that they had displayed in 25 years so since 1992, which is interesting. But something thats interesting to note here is that this is this could easily have been covered as the nefarious tactics say promoting the tories or another party. And thats something that i as a researcher want to highlight, basically something that we have advocated for the Digital Intelligence lap and the other laps i have worked for for the past five past five years that ive been thinking about these problems is the the idea that we should regulate uses of the tools but not the tools themselves. And we have to have a non partisan enforcement of how these tools are used. We can trace it in one case and say in one case targeting people with data score but when it happens with Cambridge Analytica, say its not okay. So i am running short on time and i might not talk about the type one case of collecting data and using Data Collected from bots to target voters, i will just mention that there is such a thing as travelers or spiders or scrapers, which are brought that can collect data on a certain individual or a number of individuals and that data in turn can be used to inform manual messaging or automated messaging which can be used to decrease voted for not. In conclusion, we have seen that there are cases of whats being used to increase and decrease voter turnout. And in all of these cases the bus were not transparent, crucially there were poisonous humans. And it is not evident to people interacting with them that they were interacting with an automated agent. We can expect this to increase i. Think we can expect this to be part of the fear took it off digital campaigning until we have regulation, kind of drawing a line between whats fair digital campaigning on the one hand, and what counts as digital social manipulation and behavior exploitation on the other. Its a lofty goal, its a tough red line to draw, but i think that is kind of central to all these problems, not only within politics but