WEBVTT 00:00.000 --> 00:29.000 Thank you so much. Good afternoon everyone. Thank you for being here on a Saturday at 5 p.m. 00:29.000 --> 00:36.000 Today I will present some ongoing work from my PhD. I tried to give it a bit of a title that draws attention. 00:36.000 --> 00:43.000 Let me know if you think the title is completely bogus and I will put you in the acknowledgments of my next paper. 00:43.000 --> 00:52.000 So what I study is data donation and as part of my ongoing research, I get to talk to lots of people about how they see data. 00:52.000 --> 00:56.000 Most of the time you hear things that you're probably already well aware of. 00:56.000 --> 01:02.000 Like, data is the new oil, data is like the sun, but one that I particularly like is data is the new soil. 01:02.000 --> 01:08.000 I like soil because it has something generative about it, something natural, something that we share, something that's just there. 01:08.000 --> 01:17.000 And when I put data is the new soil in Dali 3, because I also like to play with General to the AI, it came out with this beautiful picture. 01:18.000 --> 01:25.000 I find it very ironic because what I study is voluntary data sharing that people share for a public interest purpose. 01:25.000 --> 01:34.000 And as we all know how this machine was fed wasn't, you know, with the same type of data that I am trying to learn more about. 01:34.000 --> 01:36.000 So hopefully that sets the scene. 01:36.000 --> 01:43.000 But also, before we dive into the content, I would like to introduce myself a bit more. It's my first time at full stem. 01:43.000 --> 01:44.000 I first time. 01:44.000 --> 01:46.000 Oh yeah, okay fun. 01:46.000 --> 01:55.000 I'm a PhD candidate in innovation studies at the Copernicus Institute of Sustainable Development at Etrecht University in the Netherlands. 01:55.000 --> 02:00.000 And what we try to do is find out how data can help make the world a better place. 02:00.000 --> 02:10.000 And in that sense, I'm looking at data donation as a method for us all getting more engaged in science and also finding new solutions based on what we already know already have. 02:11.000 --> 02:15.000 My academic background is in economics and financial law. 02:15.000 --> 02:18.000 Actually, financial specialized in insolvency. 02:18.000 --> 02:24.000 This was about, I've always been very interested in this idea of failure and what we do after. 02:24.000 --> 02:31.000 After my master's at the study to work a bit, I worked in the construction industry working on public private financing. 02:31.000 --> 02:36.000 So we have to think of our bridges who's going to do the insurance for that, very rewarding work. 02:36.000 --> 02:41.000 If anyone was ever interested in doing something other than foster them related things. 02:41.000 --> 02:51.000 And afterwards I started teaching tax law at the University of Plight Sciences in Amsterdam where I was involved in a minor called the internet is broken and we're going to repair it. 02:51.000 --> 02:54.000 There's a bunch of different students from different disciplines. 02:54.000 --> 02:59.000 And we're just seeing, okay, what is wrong about the internet and what can we do to fix it? 02:59.000 --> 03:06.000 So returning back to this idea of, okay, repair, repair, repair, I want to do something with my PhD. 03:06.000 --> 03:09.000 I am a bit of a sentimental person I have to admit. 03:09.000 --> 03:14.000 Also today is officially the second anniversary of my PhD. 03:14.000 --> 03:16.000 I started two years ago, I'm halfway. 03:16.000 --> 03:19.000 So for me, it's also like looking back a week evening. 03:19.000 --> 03:22.000 So my laptop is also in need of repair. 03:22.000 --> 03:24.000 And I wasn't done with the slides. 03:24.000 --> 03:26.000 So I had to ask for a lot of help. 03:26.000 --> 03:27.000 Thank you, Victor. 03:27.000 --> 03:33.000 Because this is his laptop and I turned my PowerPoint, which is not free open source. 03:33.000 --> 03:38.000 So really, it is my first time at Boston into a PDF. 03:38.000 --> 03:48.000 And I also like to thank Zaz Akunitler, a good friend of mine who took the time to turn my PowerPoint into a PDF because I do not have a laptop. 03:48.000 --> 03:50.000 My laptop is in need of repair. 03:50.000 --> 03:53.000 My presentation was in need of repair. 03:53.000 --> 03:59.000 The way I'm seeing repair is really all the steps in between breakdown and the solution. 03:59.000 --> 04:03.000 But when we talk about digital repair, I would like to focus on some more common practices. 04:03.000 --> 04:06.000 Because when I was making the slides, I did not know that my laptop wouldn't work. 04:06.000 --> 04:08.000 So it's just something I'm putting in now. 04:08.000 --> 04:14.000 But you also think of Wikipedia, moderation or community-led software maintenance as sites of digital repair. 04:14.000 --> 04:16.000 Something goes wrong. 04:16.000 --> 04:20.000 People come together and try to fix it in one way or another. 04:21.000 --> 04:25.000 So how is data donation a site of repair? 04:25.000 --> 04:29.000 This is more of a working question, something I'm asking myself for this point. 04:29.000 --> 04:34.000 Because when you look at breakdown, what could breakdown mean in the case of data donation? 04:34.000 --> 04:36.000 In the case of the laptop, it is clear. 04:36.000 --> 04:39.000 It is not turning on when I put in the power charger. 04:39.000 --> 04:41.000 It doesn't come on with the right light. 04:41.000 --> 04:44.000 So it's very clear when it works and when it doesn't. 04:44.000 --> 04:48.000 When we look at data donations or the voluntary data sharing that I'm talking about, 04:49.000 --> 04:51.000 the breakdown I have identified so far. 04:51.000 --> 04:54.000 With a question mark, it's two fold. 04:54.000 --> 04:57.000 First of all, it's the lack of access when it comes to data sharing. 04:57.000 --> 05:01.000 So there are a few large pages which I will not need. 05:01.000 --> 05:08.000 Which have more control and access to data than other parties may have. 05:08.000 --> 05:11.000 And also there's a case of mistrust. 05:11.000 --> 05:15.000 Who do we share with our personal data more specifically? 05:15.000 --> 05:16.000 To repair. 05:16.000 --> 05:19.000 There are different ways to dealing with this problem. 05:19.000 --> 05:24.000 And one of them is building alternative infrastructures. 05:24.000 --> 05:28.000 And another one is to formalize trust through policy. 05:28.000 --> 05:34.000 To think of new logos, new standards, new protocols, make sure that now all of a sudden there. 05:34.000 --> 05:36.000 This was called the API age. 05:36.000 --> 05:40.000 However, due to either exorbitant costs or just straightforward, 05:40.000 --> 05:41.000 it not being allowed anymore. 05:41.000 --> 05:44.000 This API is not being shared with researchers anymore. 05:44.000 --> 05:51.000 We have now entered the post API age and we need to find different ways of accessing the same data. 05:51.000 --> 05:58.000 In the EU, we have the GDPR which really allows us all to ask for our personal data. 05:58.000 --> 06:02.000 And these researchers have gone on with it to see, okay, what can it do? 06:02.000 --> 06:06.000 What do you actually need from the law in order for our research tools to work? 06:06.000 --> 06:10.000 But also what are the alternatives to APIs if it's not data donation? 06:10.000 --> 06:13.000 So for example, screen tracking. 06:13.000 --> 06:17.000 When we look at port that's after this on their websites, very lovely. 06:17.000 --> 06:19.000 It gives you an idea of how it works. 06:19.000 --> 06:20.000 I'll just walk you through it. 06:20.000 --> 06:23.000 So first, you need informed consent of the participant. 06:23.000 --> 06:27.000 So the so called data donor that could be any of us who has an account somewhere. 06:27.000 --> 06:32.000 You can just go to the website and request your personal data and come in a nice little package. 06:32.000 --> 06:35.000 Then this data is what they call digital trace data, 06:35.000 --> 06:39.000 because everything you do online on these platforms, they just keep track of it. 06:39.000 --> 06:43.000 You first stored locally on your own device. 06:43.000 --> 06:44.000 That could be your phone. 06:44.000 --> 06:45.000 It could be your laptop. 06:45.000 --> 06:51.000 And then using port you can select which data points you would like to share with the researchers. 06:51.000 --> 06:57.000 And once you've shared it, the researcher has it and they can work with it themselves. 06:57.000 --> 07:01.000 There are some nice examples of what port can do. 07:01.000 --> 07:03.000 These are just four of them. 07:03.000 --> 07:07.000 If you go to the website data donation EU, you can find many more examples. 07:07.000 --> 07:11.000 For example, you can see what is going on when people watch Netflix. 07:11.000 --> 07:14.000 Netflix is famous for not exactly sharing their viewership numbers. 07:14.000 --> 07:16.000 It's anything but open. 07:16.000 --> 07:18.000 But if you get lots of people to share the data, 07:18.000 --> 07:22.000 we can actually get a sense of what's going on on these platforms. 07:22.000 --> 07:26.000 One that I also really like is the mapping the digital food environment using YouTube. 07:26.000 --> 07:28.000 I thought that was very original and creative. 07:28.000 --> 07:32.000 Something that didn't expect to find with data donation research. 07:32.000 --> 07:34.000 But there's a lot going on in this field. 07:34.000 --> 07:38.000 But there's also something that we can all participate in if we like to. 07:38.000 --> 07:43.000 Because I assume most of us have an account at one platform or another. 07:43.000 --> 07:50.000 So you can always check out which projects are going on and where you would like to donate your data. 07:50.000 --> 07:55.000 However, the research also identified that this isn't going perfectly. 07:55.000 --> 08:01.000 And I guess it might be particularly relevant for them to find out where the issues are when it comes to. 08:02.000 --> 08:05.000 The type of data packages that we get. 08:05.000 --> 08:10.000 So we see, for example, that most of the issues come from incomplete data packages. 08:10.000 --> 08:12.000 So some information is missing. 08:12.000 --> 08:17.000 And this directly comes from the platform provided because they are not sharing it. 08:17.000 --> 08:20.000 This might not have been an issue if we were not aware of it. 08:20.000 --> 08:30.000 So as data donation researchers are doing this work to replace a way of doing research that has been blocked by platforms. 08:30.000 --> 08:38.000 We're still finding out the platforms by not doing everything they're supposed to do players and actors to ensure there's some trust and transparency going on. 08:38.000 --> 08:41.000 This is something I mapped out in the previous paper. 08:41.000 --> 08:47.000 You can recognize the data altruism organization because they have one of those logos up there. 08:47.000 --> 08:48.000 It's very recognizable. 08:48.000 --> 08:53.000 So far, there's only one organization that has registered. 08:53.000 --> 09:01.000 But anyone can register as long as it's a non-profit that has a sole aim of being an intermediary for data for some public interest purpose. 09:01.000 --> 09:06.000 The public interest purposes have to be aligned with some national interest. 09:06.000 --> 09:13.000 But as you can see, the first one that I've turned a little circle around it is for scientific research. 09:13.000 --> 09:18.000 So this also great opportunity for those who are interested in opening up research a bit more. 09:18.000 --> 09:25.000 To make sure that it's attractive for other people and that people know that it's trustworthy and it's transparent. 09:25.000 --> 09:31.000 It also has the compliance and trust from the EU. 09:31.000 --> 09:38.000 The first one, the first data altruism organization we have is data log and no one is based in Spain. 09:38.000 --> 09:46.000 What data log does is they try to make sure that users get to know more about how they use their data. 09:46.000 --> 09:55.000 So people in Barcelona can give their household data, so energy, water consumption, all that type of data. 09:55.000 --> 10:04.000 To date a log, they in return to get a personalized report saying, okay, this is how your use differs or could be better managed. 10:04.000 --> 10:10.000 And then they use that data for all other types of scientific purposes. 10:10.000 --> 10:14.000 And the last one we have is the social economy code of conduct. 10:14.000 --> 10:20.000 So what we have here is a project that I was involved in. This is me sometimes I wear a suit. 10:20.000 --> 10:27.000 And we were working on a code of conduct together with all these different social economy actors to find out what's going on. 10:27.000 --> 10:30.000 That we did a lot of what we call articulation work within repair studies. 10:30.000 --> 10:34.000 So making very explicit what is going on, what are we trying to do? 10:34.000 --> 10:40.000 We made a list of values, we made a checklist for all these organizations and potential data donors. 10:40.000 --> 10:46.000 But also we made a list of 30 examples of best practices or organizations that are doing well. 10:46.000 --> 10:51.000 Some of them are even here today at Boston such as open food facts. 10:51.000 --> 10:54.000 So this isn't anything new. 10:54.000 --> 11:00.000 But when we were doing all of this work, we weren't talking to any developers in the previous presentation. 11:00.000 --> 11:04.000 We also have different circles and they don't really overlap. 11:04.000 --> 11:13.000 But I'm trying to say right here, there's lots going on in the EU and there are lots of participatory activities and co-creation sessions. 11:13.000 --> 11:19.000 So if there's anyone here who has any idea on how to make sure that openness is also guaranteed in the systems. 11:19.000 --> 11:23.000 Feel free to join. Feel free to talk to me after the session. 11:23.000 --> 11:26.000 And we can see what's possible. 11:27.000 --> 11:30.000 So that brings me to the end of my presentation. 11:30.000 --> 11:36.000 I have some discussion questions, but these are things that I find interesting because once again I'm working on my PhDs. 11:36.000 --> 11:38.000 If you'd like to think along, that would be great. 11:38.000 --> 11:43.000 But if you have any other questions right now, I will be very happy to answer them. 11:43.000 --> 11:44.000 Thank you. 11:44.000 --> 11:51.000 Yes. 11:51.000 --> 11:55.000 I used the early parts, like I've noticed, because I apologize. 11:55.000 --> 11:59.000 It only was a bit of a donation available to me. 11:59.000 --> 12:02.000 What should be on the common monopoly data spaces? 12:02.000 --> 12:05.000 And that's making available a lot of them. 12:05.000 --> 12:08.000 They're particularly hard and they use. 12:08.000 --> 12:17.000 And the AI also impregnages the people out of development and data storage centers to be able to use that. 12:17.000 --> 12:23.000 This done, the collaboration with you are like, how do you work with that? 12:23.000 --> 12:25.000 Repeat the question. 12:25.000 --> 12:26.000 Thank you so much. 12:26.000 --> 12:31.000 Sir was asking, what is the relationship between this whole data donation data 12:31.000 --> 12:35.000 and ongoing effort for the European data spaces? 12:35.000 --> 12:38.000 When we look at, for example, this code of conduct. 12:38.000 --> 12:40.000 There's a lot of ongoing work right now. 12:40.000 --> 12:43.000 So it isn't very clear how everything relates to each other. 12:43.000 --> 12:46.000 Actually, we're in the process of laying all of that down. 12:46.000 --> 12:49.000 So here in this document, we do try to make a link between that, 12:49.000 --> 12:55.000 especially promoting this work at the European Data Space Support Center. 12:55.000 --> 12:59.000 But because it's at this stage where lots of things are being formed, 12:59.000 --> 13:09.000 it's actually quite attractive if you have any ideas to also just send them an email. 13:09.000 --> 13:18.000 Are there any open or ongoing projects which aim to make it easier for the donation data 13:18.000 --> 13:24.000 basically to make it easier to donate the data just having an online questionnaire 13:24.000 --> 13:28.000 or what's your name, which is where you live in. 13:28.000 --> 13:37.000 Do you rent your own apartment or whatever, just for them to identify your accounts for you 13:37.000 --> 13:44.000 and then asking for you for that data and thus allowing them to automatically download 13:44.000 --> 13:48.000 based on a series of simple questions rather than going through. 13:48.000 --> 13:56.000 It's easy for us to ask each individual potential data over what the data is that they hold 13:56.000 --> 13:59.000 and then forwarding it and actually. 13:59.000 --> 14:04.000 So if I understand the question correctly, the answer is of shortened. 14:04.000 --> 14:06.000 The question is shortened. 14:06.000 --> 14:11.000 Is there a central point where a potential data donor can volunteer 14:11.000 --> 14:16.000 and then they're given a set of questions so they can be identified where their data may go? 14:16.000 --> 14:22.000 No, at this point I am not aware of any such point, but if you think that's something you could build, 14:22.000 --> 14:24.000 that will be amazing and we can talk about it. 14:24.000 --> 14:27.000 Yes? 14:27.000 --> 14:30.000 Yes? 14:30.000 --> 14:32.000 Sorry. 14:32.000 --> 14:42.000 So when it comes to data sharing a problem that I see is if you'd go for an opt-in approach 14:42.000 --> 14:46.000 that would give you a massive selection bias. 14:46.000 --> 14:55.000 I don't know how it's currently done, but I was also thinking. 14:55.000 --> 15:01.000 I think a lot of tracking uses cookies which are stored locally. 15:01.000 --> 15:09.000 So I think it would also be interesting to look at maybe having an approach 15:09.000 --> 15:18.000 where trackers that are in the form of cookies that are locally installed by the big data 15:18.000 --> 15:30.000 miners to give users the option to also send that same information centrally. 15:30.000 --> 15:41.000 So when it comes to a selection bias, is there already a discussion on whether it's opt-in or opt-out? 15:41.000 --> 15:46.000 Because I think opt-in would have a huge selection bias. 15:46.000 --> 15:52.000 So the question is, date selection if you consider an opt-in system instead of an opt-out system, 15:52.000 --> 15:56.000 how do we deal with the possible selection bias that arises? 15:56.000 --> 16:01.000 Well, when it comes to data and the ways being shared in use, especially algorithmic systems, 16:01.000 --> 16:04.000 there is already a huge bias. 16:04.000 --> 16:08.000 People from different groups of the population share the data at different rates, 16:08.000 --> 16:13.000 which unfortunately does cause adverse effects that we didn't plan beforehand. 16:13.000 --> 16:19.000 When we look at an opt-in or opt-out system, when you compare for example to organ donation, 16:19.000 --> 16:24.000 then when you go for an automatic opt-out system, that already goes much higher, 16:24.000 --> 16:29.000 however, when it comes to data, the purposes can be much wider than when it comes to organ donation. 16:29.000 --> 16:35.000 Organ can only be used at a specific place in one person's body, whereas data could be used for public interest purpose, 16:35.000 --> 16:38.000 but also a private profit driven purpose. 16:38.000 --> 16:43.000 So because of the nature of data, we cannot go for an opt-out system in the same way, 16:43.000 --> 16:47.000 but we can ask centralized servers, think of for example, 16:47.000 --> 16:51.000 internet service providers or anything else, what they do specifically, 16:51.000 --> 16:58.000 non-personal data or anonymize data to see if we can make more or better use of data. 16:58.000 --> 17:01.000 Well, what I'm proposing is... 17:01.000 --> 17:02.000 Yeah. 17:02.000 --> 17:03.000 Yeah. 17:03.000 --> 17:04.000 Yeah. 17:04.000 --> 17:05.000 Yeah. 17:05.000 --> 17:06.000 Yeah. 17:06.000 --> 17:07.000 Yeah. 17:07.000 --> 17:08.000 Yeah. 17:08.000 --> 17:09.000 Okay. 17:09.000 --> 17:10.000 Let's have a good one again. 17:17.000 --> 17:19.000 Thank you.