 Ich helfe gerne auf diesen Chaos-Veranstaltungen. Vielen Dank für das präthetische Beginn. Guten Morgen. Wir sind glücklich, dass ihr euch in solchen Nummern erweitert. Wir wollen über das Cambridge Analytica-Skandal reden. Es geht nicht um die politische Manipulation der öffentlichen Opinien. Wir wollen nach dem Jahr 2018 schauen und über das Debatte, das für das erste Mal in meiner Opinion international zu sein. Wir wollen über ein paar Ideen, die aus Brüssel und Strasbourg kommen, wie die upcoming EU-Elektionen in some pressure bringen. Wir haben die Demands, die wir denken sind, useful to avoid disasters as we have experienced in the past two years or at least so that we can recognise them as they are happening, if not prevent them. The scandal we are talking about has happened years ago or is still ongoing since a few years. We want to talk about how we can at least recognise this sort of backstabbing manipulation. We will start with a bit of a definition of terms. Just to make sure we all have present what this is about. The scandal has been a few months ago. Let's just make sure what are we talking about when we are talking about micro-targeting. We are talking about specifically talking to small and smallest groups. Groups that are political or commercial groups that are very, very targeted and talk to very directly. So, for micro-targeting you always need two things. It's not just tailoring these messages and sending these messages, but it's always the analysis that goes before then that defines these target groups and these targets. That's always based on data profiles, demographic factors, local factors or geographic factors, Economic status, behavioural data, preferences, properties and even psychometric analysis. We will talk about what that is in a bit. So this method is originally with commercial marketing but is widespread and keeps spreading in the political area. Let's just say it's not about looking at specific actors, but more about a structural review of these methods and the toolkit that exists to do so. And we are not going to talk about effects mainly. Empirical studies on how micro-targeting works are rare and there are very different and controversial interpretations of this phenomenon and we'll just leave it at it's very context specific as to how it works and it really depends on which actors use it for what and in what scope. And if we want to keep things simple, there is a bit of an analysis before these campaigns start and in our ... Oh, and there's been a talk about police laws and their capabilities and so on. And as a party wants to support that or be against it, they do analysis of voters and maybe the analyst will say, well, as of this age of water, this message will be particularly effective and so they try to really tailor these messages to each other. And I believe that it's important to stress that it's really targeted, very, very specific and we will talk about this into more detail about what Friederika talked about yesterday from Privacy International and it's also about people who actually do not have an account on these platforms and that people actually are being analysed across several devices that they own and so let's say if you are just sitting on the tram and reading news on the way to work, then it will still be assigned and recognised as you. In the year 2018, the whole thing started but the fake news problem was actually much, much older than that and we would like to ... So we really want to talk about ... ... not talk about fake news but we want to talk about the targeting aspect in this very talk and maybe some people of you might remember if you still had Russian in school and of course before the digital era people were trying to influence the general population with propaganda and of course with the technical possibilities today we have a completely different approach that is still in principle the same as in the older days and it's really important to remember that there is a lot of experience and old mythology that is creating the foundation and also creating the understanding on part of the people who try to manipulate and if you think about manipulations not only during the Trump political campaigns but also during the Obama era people were always trying to influence other people and trying to influence the political public opinion and they were really interested in making everything functional, efficient and effective and so there is no final result yet some of you might have read the papers that investigated these issues but we still don't know that much or we still don't have final conclusions so there are some investigations going on but there is no final conclusion yet for these investigations and again to the original setting so if you think about the European Union election next year so how is the current political situation for once it's we see that a great of nervousness in the political public and especially in the digital public there is a lot of mistrust against established parties but also a general distrust against all news on the internet and the second aspect is that there are basically two big companies that are mostly responsible for digital publicity and that's just Facebook okay so these don't just produce content on their own side but of course on the other side and Google with YouTube and the double click ad network or analytics services they are extremely important in terms of creating a digital public space and we can see both of these services providers are coming and they are an integral way of how digital public is structured and created and that's the inside tooling so that that provide people who have websites and run these campaigns know what they see and there are toolkits for this sort of ad targeting where you can very easily put together what does my target group look like where are they from how old are they, what properties do they have and it's easy to do that and a lot of people kind of know that but it's sort of not very present in public debate is that both Facebook and Google there is a custom audience function which allows advertisers to contact people that they've been in touch before whether you're a political or commercial advertiser if you already have contact data emails, phone numbers then I can upload those to the services or check with these people and reach the people I want to reach and moving on from there there is the lookalike audience function where I don't just get access to the people who I already have contact data of but also people who are the same or very similar as far as their data profiles go so that I already have in my database and so all of these tools that are already provided by these platforms are an integral part of this micro targeting without which none of this digital public would work so now that the GDPR has gone in force that had some consequences if I have a mailing list already am I even still allowed to upload that to these services and there was some discussion about that so in Great Britain all of the political parties have been requested not to do that and that will be done in the courts and in Germany we have this discussion as well and the Bavarian Data Protection Agency has the same view and will try to enforce that as much as they can well in the Cambridge Analytica scandal it was interesting the questions that came up in the investigations for the Cambridge Analytica scandal were mostly interesting because these companies finally had to answer some questions so was interesting there was that you could see what was liked what ads did people click on what apps do people use if you have people who use these services and then you can get data that cannot be used due to this discrimination prohibition and well this stupid Facebook well there are two types of scandals and we probably forgot what the essence of this scandal is why it was so big on an international level and then also the parlaments basically said okay we are going to get back to you on these issues so even before the scandal there were reports on scandals with Cambridge Analyticas but not in that magnitude this scandal is a special one because first of all it's about Facebook and they are known for being not very strict about privacy and secondly because data has been used without user knowledge and without user consent for this targeting in political campaigns so let's quickly summarize for what was actually the issue with Cambridge Analytica so there was how was it founded there was a scientist that basically created then Cambridge Analytica and there are actually some interesting name interesting stories how the name Cambridge Analytica came to be maybe it's because so okay not by the translator I couldn't understand that unfortunately and so so what they used was an app that was originally meant to be used for scientific analysis and they used that they hired these click workers working for sense they took that app which connected to Facebook and filled out this questionnaire according to the ocean model which is an established psychological model so that these people could be sort of their personalities could be determined by these five characteristics so as they filled that out they provided them with a very rich data set from Facebook including all of their friends and that came up to I think 80.000 if I heard that correctly people so they took these personality analysis of the 23.000 who filled out the questionnaires and searched for similarities amongst the data profiles on Facebook so that they ended up with psychological Sessionality Problems of thousands or even millions if I heard that correctly people had and so in the US they had a lot of more relaxed access to voter data and so they used all of that data to try and support the Trump campaign in their political communication so as Christopher Riley the whistleblower describes they used that to create these fitting ecosystems of information including blogs and other websites so that they could target people and they created this impression of a neutral public and neutral reporting by creating these websites and blogs that they could then show to people on topics from gun laws to immigration and of course that was mostly on political issues that were very polarising and divisive and of course they used it for micro targeting in door to door campaigning which is a very important thing in the US and they created up to 33 conversation guidelines and they knew if I ring the doorbell on this door which guidelines should I use and they used it even on Hillary voters because they knew that certain voter groups would not be reachable for that to strategically demobilise the opponents so in regions where Hillary is strong or would be historically particularly strong women people of colour they specifically sent them negative information to discredit that candidate so women got a lot of advertising which talked about the affair of Hillary Clinton to undermine the image of Hillary Clinton as this brave feminist that the democratic party was painting and that's sort of a lot more widespread than in European or German political advertising so there was a lot more possibility for manipulation there because this character advertising is much more important there is no definitive analysis how much this affected the result it's clear that this manipulation happened and that it would have some effect but we can't say definitively if there wouldn't have been Donald Trump without this we do know that if we know about the distribution of the voters we can do our political advertising much more effectively and again we can use these demotivating messages and of course that played a much bigger role than it did in the Obama campaign which did a lot of that too but was much less criticized for it and of course it boiled down to 27.000 votes in three different states so it was a really really tight race at the end so we also want to talk about the Brexit referendum because that was also terribly terribly terribly close and because in the British Parliament the debates the investigations are a lot more public and a lot more precise and the statements that that happened afterwards in British Parliament they brought entire catalogs of questions for the corporations and Cambridge Analytica and about what is exactly happening and being done and we can always just talk about all of these different companies and data companies but we also have to look at the political parties who regulate this and how much are they interested in that themselves how big is the conflict of interest there since I mean we are clearly talking about questions of power and after like in the US every four years somebody is elected and Trump could be president for eight years but Brexit that will be a problem for generations and it will be much more important question and much longer lasting and it was even more close so we can see with the Brexit referendum that the pro Brexit side leave exchanged a lot of data internally between the different groups there which is not exactly appropriate following the rules on data sharing just because I provide my data to one campaign I did not consent to them sharing this data with other campaigns and other corporations that are formally independent and not supposed to work together and we also saw that the leadership of one of these of an insurance company was very close to one of the pro leave campaign and so the customer data of that insurance company was used to target people for pro Brexit campaigns and it's also important to note that leave also used a lot of question a lot of messages of questionable veracity and outright lies and the investigations on that are still not done ok as we are running out of time we are keeping the next few topics short but we want to mention them in Ireland there was a referendum about abortion about the law about abortion which was very strict and it was a very polarised debate in the public and among other things we could notice that in this political debate there was not only people Ireland influencing the debate but also religious fundamentalists from the United States also tried to influence the outcome of these debates and they actually set up websites in the United States to influence the general public's opinion about this debate and so to highlight this was just about very low sums and and people were actually spending a good amount of money on funding, campaigns even if it didn't affect their own country and of course this is something you don't notice until the whole campaign and the whole debate is over another example is the election in Kenya in 2017 there were two parties and what we could see is that in Kenya and in many other African countries is that privacy laws are a lot weaker and all the parties have a huge amount of data about people including biometric data, telephone numbers we also saw a lot of negative campaigning against with the campaigns of the former and re-elected president by using by using negative campaigning and what we've seen is that a development that somewhat foreshadows what we will see in Europe as well that there's a lot of things going on with messengers so for instance WhatsApp groups or other messengers that people were really targeting people on these messaging channels and this was easy because the parties knew the telephone numbers of those people that's why they could target them and we already know that 2019 there will be a lot of elections and it's it's very important and important and interesting to look at the new communication channels but but what we should what we should also do is like follow the money see who spends money where on which watch and and we will probably see many new technical strategies to influence people and of course if you consider Facebook different continents have a different usage of Facebook and yeah for instance in France you had a great debate about okay so now the election in 2017 and I hope you will see a pattern in Austria in the election in Austria we had seen that an advisor and election manager of a social party was was sehr well worth with these methods he used faked facebook pages so he created a website on facebook called we for the bastion courts and and they looked like it's all real but this was meant to okay notifying me of the time that's running out so it was really like misinforming people it had a lot of followers and it was also used to use the data to know who could be targeted with based on the people who left data on that facebook website of course as you can imagine that was some fierce debates here and obviously it's still not clear if these attempted campaigns really affected the results but it's I mean we have to see that in Austria this microtargeting and negative campaigning has been a lot more strong even if we look at other examples than the this most crazy example of Tal Silverstein in Germany um there was a lot of nervousness there a lot of fear especially after the Bundestag hack and that these there would be leaks from the Bundestags hack so there was a lot of nervousness and insecurity ahead of the election and the party is a part of a rather very vague self declaration didn't really do anything to get rid of this fear and insecurity at least not by providing transparency us at netspolitik.org we tried to create some transparency there like how much money is being spent and so on but we basically we ran into a wall there we straight up asked them and the ones who regulate these things so we just went and asked them show us who are you paying what for you are creating these disinformation campaigns what's the point and the responses were pretty much zero Green Party were the most open Die Lincoln the leftist party talked about it fairly openly but all the others they did not answer they did very little to create transparency and information so since we don't want to talk about the specific actors or examples but rather about methods let's summarize what are the methods we have to watch out for what do we have to expect in 2019 so we tried just writing this so seduction messages tailored to individual personality profiles we don't want to say that like that this is the responsibility of the individual voter or anything but we all to be aware of these methods of attempted manipulation of course it can't be everyone's individual responsibility we need to have a political response and information this list may help we're not going to go through all of them one by one let's talk about the measures that have already been taken in the EU so since December new there is this this catalogue of measures that shall be taken that comes from the EU commission so that's the first time they actually did something at all and that was all they did EU regulation does not happen fast so they put down some proposals and requested all of the member countries to implement them and that's more than just this information but well it's kind of a big thing okay so they don't just want to try to stop this information but they want to provide and pay for counter propaganda and the British company British government ran into trouble here they had problems because a group that they paid that was supposed to work against Russian propaganda suddenly was targeting the leader of the opposition Jeremy Corbyn and now we are building the same structures for counter propaganda that didn't work in a couple of countries and in Britain that actually had quite some problems and it's going to have consequences in other countries as well because they targeted Spain because their the head of their secret service wasn't they weren't happy with that person so they targeted Spain so that's going to have some consequences and it's that's not a good situation but a lot of parties are not going to care because they want to use these methods themselves so but we have to say also that the EU doesn't really have the competences in this area it can't do a lot it's a bit late but they are trying to work on some monitoring of these methods and this information they they had to recognize early 2018 that the election commissions in the various countries just didn't know some of them just didn't answer the questions that the EU asked them so not just that they weren't ready, didn't have any measures in place or prepared they just didn't react even so the EU parliament passed some demands but they are not binding so we'll pass over those and go on to our demands so the first one we had this polarizing debate about the GDPR we believe that our first demand which is data protection is manipulation protection because as individual countries in Europe start to recognize that political manipulation is done with these with this data that is being gathered then it becomes very pressing that data protection laws are in place here if only so we can see if people are trying to do this manipulation and at the same time data protection is manipulation protection and that that even that applies especially also to shadow profiles so even if you aren't using it data was gathered about you and so that becomes ever more paramount we do have some specific questions here well we're gonna try to go through this in the last two minutes we got our sign signal now political campaigning is always about convincing people but what we can see is that these digital platforms with their tools make it possible to have this really really strong asymmetry between those who do political communication and campaigning and those who are being manipulated so it's absolutely necessary to create transparency there so that political advertising is labeled and that you can always see at least with the possibility to see who is targeting me who is paying for this ad and of course that's not just a transparency register but a collection of information on which providers actually provide this ability who are they and who uses them with what money the British actually gathered this information and you can look at the data from 2018 to 2018 you can see where that money is going and you can see you know is cross devising being used and which corporation gets the money 3.2 million pounds were used for that and which isn't very little money and we just don't have that data in Germany for example we should change that but that would mean that we have to change the party law in Germany as in there where the main rules are for transparency for political parties and those are okay but they aren't about where that money goes and to create this sort of equality of possibilities for candidates we have that for TV but we don't have that for these advertising companies so we don't have this equality of opportunity for political candidates for these advertising companies that's one but of course we as voters have to take these political parties to task we have to know what they are using the money for with which criteria are they targeting people and it's not okay that we have to go after that information and keep asking and begging them for that information so Psychometric profiling is already tricky and really that's targeting I wouldn't be sad if it were forbidden to target particularly vulnerable segments of society So Facebook und Google create this manipulation force that is so big that we cannot deal with this with smaller measures. Smaller measures are nice and all, but we really have to take big steps to deal with these tools that they just hand out, which have such a great potential and we have to regulate that and deal with that. Sorry, it's our last sentence, really, we'll write about this again and talk about all the things we skipped. Übrigens sind wir der Meinung, dass Facebook und Google es sind.