 Gwyddech chi'n gweithio'r srpan a'r gweithio'r gweithio i'r Gweithio'r Sleidio, yw'n gweld i'r gweithio'r ysgol i gweithio'r gweithio, ond y gallwn i gweithio'r gweithio'r srpan i gweithio'r srpan, mae'n siarad, yn ôl yma'n ymlaen. Gemma, maen nhw'n gymryd? Rhaid, dweud, mae'n synod i'r unig a'n gyflaen i'r Gwladol. I specifically work in a department that provides advisory services to governments in Africa. I'm part of a team now looking at what kind of tools and innovative ideas that we can incorporate in our work. I'm here just to learn, but one of the challenges is all the presentations I've seen have been incredibly interesting and enlightening and the kind of thing that makes me think, oh yeah, let me try and let's see if we can pilot this somewhere. One of the challenges I have is where does the money come from to create bespoke versions or to apply some of these tools because where that money comes from can sometimes undermine the integrity of its application. So I'd be curious to hear in terms of like a practical take away for people like me who are really keen to introduce some of these tools and technologies in developing contexts, who should fund it, where should the money come from and can we do it for free and what are the risks of doing it for free, thanks. There we go, hello. I'm probably just because of the area we work in best suited to answer this and my answer won't be helpful because in truth I don't know and I guess what I'm trying to say is in ways of keeping your operation free of bias and free of influence from where the money comes from is maybe protecting the people who actually do the work from that so I don't necessarily know when we're working in Nigeria or Kenya who exactly is paying me and that's not necessarily a bad thing. If I don't know who's paying me then I'm somewhat protected from the bias and from the influence of that money. I know that doesn't answer your question where to get the money from but... I would say that a lot of people at this conference would be able to address that though so do you keep asking the question and don't feel too stupid to ask any question. What they said. Looks like we have some questions on the left-hand side of the room. Hi, my name is Sohar Yadagari from Adesium Foundation which is a Dutch flantrape organisation. My question is to the last speaker. I'm assuming that one of the main factors why this type of open dialogue could work is basically the feedback loop is completed. So when I share something with a parliamentarian or representative I get feedback that that person is actually doing something with that information and I also experience this in real life. Is that a correct assumption? And to what extent is that really the case? And do people not become completely disappointed in their act of participation? Good question. So yeah the concept definitely is this idea that you could raise an issue through something like a WhatsApp type channel and then on that so then an AI would be kind of in the middle of this conversation so you're sending something to your representative and either the representative does act on it or doesn't act on it but I guess the computer in the middle of this conversation is updating you in a non-biased way and that is kind of keep feeding you information. The concept here is that rather than you going to the newspaper and opening up and flicking through these stories that some of them are of interest to you and some of them are not of interest to you that you're actually getting direct hits on topics that you have actually engaged on from what is I guess a probably government funded source I'm not too sure but in some places on the topic that you've already engaged on and it's meant to be kind of non-biased like here is what happened and you can go elsewhere to find political commentary on that and so that feedback loop then I think is something that's very we're passionate about and we showed in the pilot study with our Cork what was not a perfect sample of the actual scenario because it was something that was actually it was based on most of the issues they receive are obviously are city issues and not policy issues so the issues you can physically see a pothole fixed or something like that but those who did receive feedback were far more likely not just to think that their voice was heard on that occasion but actually to think that their place in all of Irish society was, I think the question there but it was basically they were forward specifically engaged in general and felt that they had a stronger voice that was in Irish society which I think if we can create those kind of changes even if it's perception that's very powerful for citizens Hi, I'm John Roussi from Transparence International my question is also for Johnny if I understood this correctly this platform would allow people to communicate with their elected representatives and their elected representatives can then respond to a lot of people at the same time so that saves time Do you have any views on whether people should also be allowed to see what other people are saying to a representative and communicate with each other because as we know group norms play a big role if you know that a lot of other people are concerned about a local school being built or a lot of people are concerned about this new law or this other thing you might also be more likely to take action on it so if the interaction that you've designed is only from one individual to their representative it doesn't capitalize on the collective action opportunities So, good question again and one of the things that we wanted to move away from a little bit was I guess that change.org type thing where you need to create a narrative and build public mass appeal for your idea so a lot of those times if you have those type of appeals you have to create a black and white type argument and win the hearts and minds of people and really build a campaign to get critical mass to get your issue kind of front and centre of public representatives but in the paper the last principle is actually kind of around that transparency and part of it is ensuring that this is basically a new form of lobbying in a way and that there is means that you can actually see what politicians are hearing on a kind of macro level and so if this was done by a platform like Facebook they may build that into the front end to encourage people to engage but in the situation that we have the concept that we've just outlined there that example it's actually kind of it's something that you can kick into where you can see what public representatives are hearing from their constituents or from other constituents so that has been built in Yes, thank you. Margo Lord from Estonia, citizen OS I have a question for Felipe The examples you were using in your presentation were sort of snapshots of when you're releasing a certain amount of data or a certain data set what is the acceptable level of anonymity that you need to create our platform is being used by people for citizen initiatives or citizen petitions that are then given to the parliament and the thing where Estonia is perhaps slightly different is that people are identified by the electronic ID online so you know for certain who voted for the petition who voted against the petition and what kind of petitions people are voting for and this means that some organizations have raised the concern with us and I think it's a legitimate concern that over time if you had access to this data you would be able to profile people based on their political or social concerns on the one hand we want to release our data for researchers as open data because a lot of the work that we're hearing here in this conference is based on platforms like ours generating this data that researchers can then use my question is if we're releasing a set of data today and making it anonymous to the required level but then we release more data the next month and the month after that and the year after that and the data kind of builds up are there ways or approaches that either this API you introduced or other tools can be used to make sure that even over time you cannot profile specific people that even if today the data released was anonymous to a certain level then after two or three or four further data releases you wouldn't be able to narrow down the groups to start profiling people. Do you do? I love the question now the answer is hard like mathematically it's super hard to preserve anonymity and that's why part of the... without knowing the answers I love providing the thoughts and tools to start thinking about that and how hard it is like yes everyone agreed that we should do KAnonymity 5 and then we are fine but we're not that's why we developed diversity KMAP and all these metrics but you are working against adversaries that the more data they collect they can find out ways to the anonymize so it's super, super hard and you need to live with some degree that we did the most we could while still providing a valuable service we need the data for valuable purposes but how do we become future proof? It's a hard mathematical problem like it doesn't matter how much we talk about it at the end of the day there are mathematical ways that you could say this is a very, very hard problem to solve but yes, we are in this together We have a question on the aisle Hi, it's Chloe Messenger from DAI Martin and John, you both mentioned how important designing with the user is and I was just wondering how far your organizations go with that whether you hold incubator style sessions or whether Martin, you put up a picture of people going to the field and talking to people and also how far you think it's necessary to go how much you need to talk to people one on one and how much you can just get the data from research Thanks Well, in our field actually we're actually a user research company we're a user experience and so most of that would be research driven from user interviews so data is usually a secondary information source so in our field, in our consultancy it would actually be the primary kind of collection of data and then we would review usage of tools if it was a digital tool we might use Google Analytics to monitor that but it's always best to have that kind of qualitative information and that would be kind of core to our principle Typically at my society I don't very often start new things but when we do start a new thing talking to people one to one as soon as possible is probably usually the best place to start as we work quite iteratively over existing products we don't always speak to people directly but we do use things like surveys analytics, lots of A-B testing I don't know if you've come across that before where you do one version of something present to 50% of people on the live site you don't set up a fake test I love those questions where you get to say it depends because it depends on exactly the problem you're trying to answer Looks like This person's first and then your second Thanks Anish Darabbi from A Political This question is for Felipe I just wanted to ask for one of your examples healthcare is one of the areas where big data analysis and on top of that AI could be the most useful and have the most transformative effect but at the same time it seems the number one in these places where anonymity is one of the has the highest stakes and where people care the most as in do you see so first of all is this what is holding this kind of advance back in healthcare and also is there a way around that problem convincing people that the incredibly personal data will be held safely and or is this a very similar problem to the previous one is it a tricky one to actually deal with That's something I think about a lot again I don't have the answers my lawyer tells me to always say that but I was talking to a researcher in imaging they were doing their PhD basically if you are at a university doing research on how to fix breast cancer you don't have data they have very few data points available and that's a huge cost for our society so we need I think we need to make this more clear that yes we are protecting our privacy which is good but on the other hand if these researchers had more data available they could advance years forward just by making more data available and coming back to my favorite example is matches do we need to make matches disappear can we make matches more safer or it's just about regulating how we do things like we can have dangerous things in our life and the problem is not how dangerous they are but yeah don't start a fire and that's regulated outside of what a match cannot do by itself and I think a lot of the solutions have to come from regulation, from ethics of what you can do or what you cannot do with data instead of limiting that itself thanks one is a quick comment for the lady who asked about whether completing the loop results in genuine impact just before I came I read a really good report called Civic Tech in the Global South Assessing Technology for the Public Good and I really recommend it and there's an interesting analysis around there are some context in which people don't care that they fail to deliver on what their promises because they're going to get voted in anyway I won't make any comment about which places they are I have a question for I've forgotten your name John So at the moment I'm part of a team that's looking at how human-centered design can be incorporated in our work in African governments and specifically looking at piloting human-centered design research and I'm very conscious of the fact that whether I like it or not I'm going to be part of an elite and that there is a certain association with government and I wondered if you had any recommendations or advice or some kind of framework looking at what can ethically guide the way that we do this research and that is respectful to vulnerable communities but also just the inevitable power dynamic that comes from being an outsider doing research That's tough and there's a number for researching this we ended up because we went to different franchise people so we went to the worst areas of Dublin went to the pub and just kind of actually just had chats we weren't writing things notes down we just kind of wanted to bring up topics and explore them which is a very slow way of kind of doing what you need to do One of the best ways of doing it though and a bit of a bigger picture and outside of Ireland is to find someone who is in the community who understands your plight and then research with them and then ask them to be kind of I guess a referral and start using their networks so you have in whichever community it is that you need to be talking to and you then have the person you have that strength so people feel more confident that they're from somebody and that person has some sort of of clout so the first creating the relationship with that initial person is important and that's typically quite a good way in terms of starting to build trust another way of doing it is to get someone from that community so we do a lot of research in different countries most of our research is not as sensitive as what you guys are doing but we would have someone maybe from that country do the research research from that country lead the research just that it means that the language buyers are broken and I guess they feel more comfortable or confident speaking to that person especially beyond kind of the surface level detail that they want to share so I think those would be two things that I would point out in terms of breaking that gap which often is a tough gap to break