The following are the outputs of the real-time captioning taken during the Fourteenth Annual Meeting of the Internet Governance Forum (IGF) in Berlin, Germany, from 25 to 29 November 2019. Although it is largely accurate, in some cases it may be incomplete or inaccurate due to inaudible passages or transcription errors. It is posted as an aid to understanding the proceedings at the event, but should not be treated as an authoritative record.
***
>> MODERATOR: Hello. I think we will get started in two minutes so if we could get the panelists on stage and other participants in the room, please.
>> MODERATOR: Hello, welcome to this highly ref vent panel on data driven business models. I will not make a long introduction because we have lost a lot of time. This panel has gone from 60 to 50 to now 40 minutes, but fortunately there has not been a reduction in the number of panelists so we are still nine panelists with about 40 minutes to share. We all know the importance of the subject. I'm not going to go on about that. Data really is the oil of this century, but we have a lot of questions that we need to discuss. How are we going to leverage that? How are we going to leverage the value of data? Data is not homogeneous, there is public and non‑public, user and machine‑derived data, so we need to discuss how we are going to leverage the data in the future.
We have done a lot of work in the OECD and we have new work now on how to enhance access to data. That is why we take a particular interest as part of our going digital project which is the biggest horizontal project in the OECD. It's going to work this way. I'm going to give all nine panelists the floor to say whatever they want about the subject. I'm not sure we are going to have time for Q and A. We will try to fee if we can. I have a method to discipline you that the public doesn't know about, but there is a watch to limit to three minutes. I can see it, but you can't see it, but that is meant to discipline us.
You all have three minutes except you Martin, you will kick us off, you have five minutes. Martin is the founder of Bolt and also an entrepreneur for 20 years. You have started six companies and you are the cofounder of the foundation. Martin the, the floor is yours.
>> MARTIN VILLIG: Hello. Great to be here today. I'll try to give a quick overview of how we see transportation and how data is relevant to our business model. Bolt is the third fastest growing company in Europe based on financial times. We operate in 34 markets. We operate micro mobility, electric scooters, motorcycle, taxis and now recently now also food delivery. We have been growing really fast. We are a 6‑year‑old company and one of the European unicorns.
How much we actually and gather data. We do about one million rides per day. So we get hundreds of gigabytes of data. When talking about AI. We have 300AI models in production currently. That's not just developing the models, but we need to monitor them to make sure they make sense, add value and so on. Then what are the use cases where we actually use data and also data science and machine learning.
So as in ride hailing, you probably know that pricing is really important. So passengers always want the cheaper price. The drivers always want to have higher price. How do we final reasonable balance in between them to make sure that we have enough drives to cover the need and the price is cheap enough so that drivers will be willing to come out. So this is one of the main things we use data and try to predict what kind of pricing logic we should provide.
Second is about campaigns.
So, for example, recently we launched in London and we invested millions per month, so how we make sure that all of the campaign money is exactly targeted to people that will get the best return out of it. So we need to subsidize drivers to bring them online. We need to give discounts to passengers, how exactly, what patterns is needed to get the best results. So this is on the pricing side. Then maps and ETAs, you probably all know about ETA of these apps are challenging.
Sometimes they go wrong because the city traffic is very difficult to predict. So we have now lost two years built the mapping layer in‑house. Before that we were based on Google Maps, but now we also use open street maps, we use our own data from the rights and we try to calculate the best possibility to make sure that we can tell you when the car actually arrives. So those are maybe the three main areas where we use data science.
A few principles, we always want to be ethical. So we don't want to use data, for example, we know that passengers ordering a car, they have an expensive latest iPhone model and their battery is about to die. Quite a good situation to put high prices but we don't want to do that, and all of our majority of our AI decisions are actually based on aggregated data, not about single person.
Few examples or a few exceptions when we need to deal with fraud prevention or law enforcement situations, then we need to consider specific users. And, of course, we need to take very much care about the data. Everyone is accessing private data is logged and only per need basis. And then lately we have lots of regulations coming up, GDPR, but also the AI ethics and different frameworks that we are monitoring on. So we try to always follow the regulations and best practices.
And then what challenges we see ahead. So for our case, operating in cities then many, more and more cities are asking data from the platform. The challenge for us aAPI is very different, how they ask, what they ask is very different, so we would be happy if there would be some kind of best practice globally so we could have one API and share it with all, but on the second the challenge is what data to share because data is also competitive advantage for platforms and cities often ask too much data which they may not even use. So we need to find a very balanced situation, what to share in order not to lose competitive advantage but on the other hand gain data that cities need for planning transport and so on.
And secondly, we see that over our global lit rasely level for people is relevantly low. We have sophisticated discussions about AI, the data and so on but most people don't understand that. So we need to raise and educate people. There is one very good finish initiative. It's a free online course, elements of AI and Finland claims that 2% of their population have already passed it. I think that many other nations could take that lead and then try to educate their people so we actually have relevant and reasonable discussions.
And finally, I mentioned radio regulations. That's a challenge everywhere, so how to find a decent balance for that. So thank you. That was a very short from Bolt, and all of our rides in Europe are carbon neutral so we also try to be a sustainable platform.
Thank you.
>> MODERATOR: Thank you very much, Martin, you did it in four minutes. That shows the efficiency of the private sector. I would like to turn to you, Dr. Dirk Abendroth. You are the chief technology officer of continental automotive and award winning experiment in the development of systems for automated driving, connected mobility and electric vehicle drivers. You have the floor.
>> DIRK ABENDROTH: Just reflecting on a couple of critical principles we think are the key ones to make business happen and make us feel comfortable with this. I think one of the key things first of all is that we agree on an underlying principle which is very simply speaking, we are under control for data. It applies to us as human beings, our personal data on the one side, and secondly, as he mentioned as well, to IP, so IP is to a certain degree data as well, and it is something that needs to be under control, otherwise no business is possible as you mentioned earlier. So that is kind of the to my understanding underlying principle otherwise it can't work.
Once this is in place and you are under control of your data, the second step is very straight forward, then it needs trust to give your data to somebody else. So this trust is something you need to earn one by one or you need to have at least enough initial trust and then see if trust is proven to be right. So in the end, it needs very clear regulations and how do I give away my IP and get recharge back. How can I have control of personal data and make sure it's no misused, et cetera.
So these are the two steps which are think are straight forward. When it comes to the very special and dedicated business I'm in automotive and mobility. I think there is three basic notions I would like to mention here. One is classification. So to us, it's very, very important what kind of data we are looking at. Just to give you a couple of examples, one, you just collect traces or another thing you might ask people and the quality of what you get back, the samples are totally different, which is related to a second aspect, which is what I call qualification. So qualification is something like, well, you could simply kind of translate that to quality. What is the quality of data I get.
So, for example, is it a representative sample is it something that represents European Union or African people or the entire world or a city, whatever? Is there a representative sample? Or is it maybe something which even qualifies, for example, to get a certain location to get something proven to be safety relevant or even safe. So these kind of traces, once they are kind of harmonized and have been proven by an official institute or Government could be very helpful for us.
So another example for qualification of data. The very last one obviously mentioned by, I guess, have people people today already by Mr. Ailed mire as awell, cybersecurity, you and I know it's easy to manipulate data as well as software so it's one of the crucial parts we need to take care of as well.
>> MODERATOR: Thank you so much, Dr. Dirk Abendroth. I would I would like to introduce, and we will jump ahead a little bit in the queue, I would like to introduce Carina Röllig, you are the founder and CEO of web data solutions which is one of the fastest growing German startups and has released market analysis of software black beak.
>> CARINA RÖLLIG: We are a spinoff of a university and it's all about data. We collect online data prices and product offers from online marketplaces and retailers and for us it's really relevant what we are allowed to do with. So first of all, for us, all of the regulations with the Internet, all things that doesn't allow us to extract product information from the Web like images or data descriptions, product descriptions.
This is everything what maybe could bring us and travel with our business model. So for us it would be very, very relevant to get the same regulations as you maybe have in the U.S. or in other states in the world. So huge differences if you compare this with the German law. This is maybe the first point. So on the other hand, as we are extracting all of these heterogenous data from global web sources, we have a very high focus on data quality. So data quality in the end is really key.
And what we see with even huge companies are sometimes struggling with data quality even with the internal data. So if we bring data together to make analysis to realize maybe in the end Artificial Intelligence solutions, we need to make sure that we can trust in the data we get, and what our clients can also rely on this data.
What we also see is what data is stored in different resources, platforms or databases, and to realize global data‑driven businesses, you need to bring all of this data together, normalize it and enhance it maybe with more relevant information, and to make sure that it's always available and it's directly usable. And for us, the main point with the Internet is on the other hand, this is really a huge duty to use the data there, and to make it useful and to bring it in the end to solutions which can help people.
Maybe in our kind it's an easy good solutions how to set your own prices in maybe your own store, but it could also help to realize which products are maybe sustainable or not. So there are lots more business cases behind this data we are collecting today. I really hope that we can create a world. The data is available and it's useful for everyone in the end.
>> MODERATOR: Thank you so much, that was short and sweet and to the point.
I now turn to. You are also a former entrepreneur, with digital experience across the public and private sector. I notice that I forgot to introduce myself earlier, and I think the easy way to do it is by saying whether Henri and I met earlier today and we talked about negotiations. He said I have been a tech guy all my life, I have only done this diplomatic stuff for the year. And for me it's the opposite, I have been a diplomat all of may life and I have been doing tech stuff for a year. Enough about me, Henri, you have the floor.
>> HENRI VERDIER: Thank you, for four minutes, and I will leave before the end of the Conference.
>> MODERATOR: But not before the end of the four minutes.
>> HENRI VERDIER: Good morning. I'm happy to share ideas with you. I was the first chief data officer of the French Government and I share four ideas. The main idea is we have to face the fact, think about this, through history the man's greatest growth has occurred when scare resources became abundant. And now we have to face the fourth one is the data abundance. We are not prepared for this. Because when we have this kind of revolution, everything changes.
The balance of power changes, the value changes and the public policies and role of the state changes. And this revolution is not like the former revolution because there is a great difference. Data is not the new oil. I disagree with your introduction, because data is an infinite resource, you can use the data and you create more data, not less data, so we have to deal with abundance revolution with an infinite resource, and that's something very new in human history. And your question this morning is how to create more value in this field.
So first, the first answer is quite simple. We create value when someone uses the data to build something. That's quite simple. And there is a first consequence is that very often, maybe not every time, but very often you should share your data, make open Government data for Government because if you don't, the people will find the data elsewhere, and will organize themselves without you. And without your data and you will weak and alone somewhere. So it's not so difficult. If some people use the data, they may create value.
The question for Government is more where do we create value and for whom? And, of course, as Government I think that we have to promote shared value and to fight for general interests and for this, maybe it was your question we will need some regulation, sometimes, but I feel that we need more public policies. We need three very important topics in these public policies.
First, we need to first have innovation ‑‑ foster innovation,, this innovation comes from a rich ecosystem with a creativity, iteration, fair competition, we need research, experimentation, we need to be able to create a lot of startups. So we have to foster innovation. Then it is very important we need to build data trucks. We spend a lot of money for roads rails, and a lot of physical infrastructures and we do not consider data infrastructures.
So sometimes we will have to finance and build infrastructures, but in this field, sometimes it's more efficient to cooperate with Wikipedia, because users, people, you are building very interesting infrastructures, digital comments is very important. And, of course, last but not least, we need to build trust and accountability into everything we do with data. Most people have concerns around the misuse of data to oppress and control individuals. They have concern with policies, with privacy. They are right. You have to face this. You have to organize our severals to avoid this kind of threat with data.
So the citizen has the right to have this consent and the Government has the duty to find solution with you, but that's my three topics, innovation, infrastructure and trust and accountability. Thank you.
>> MODERATOR: Thank you so much. I would like to turn to Grégoire Kopp. You are the Chief of Staff at OVH, you have a variousiad background as former Director of communication and spokesperson for Uber France and you are also previous ministerial advisor in the Minister of Transport, and you did some time as a lawyer, lobbyist protecting consumer interest. The floor is yours.
>> GRÉGOIRE KOPP: Thank you for inviting us to talk. It's good to have area to dispose between diplomats, politicians abecause we need to create Internet together. OVH cloud company is the biggest one in Europe. We are the only European one in the top ten worldwide and we were built 20 years ago by a polish immigrant. So we are open by design, we create the European scale between 2000 and 2010 and we did that with very special values. Because the founder is a real tech guy, so for him, open source, sharing data is very important, and all of the commitment of the company is based on that.
First of all the motto of the company is innovation is freedom because for us it is natural that innovation can create freedom, but we discovered and realized a fuse years ago that it was not always the case, and we change our motto to innovation for freedom because we need a special commitment for that.
To answer the question how we can build more trust for Internet based on data, there are two main points. The first is reversibility, we need to exchange data and we need to create some way to work together. Now, in Europe there is a special initiative based on the German Government because we need to create links between our companies and we need to share it with transparency, and we need to push standardization too. That's my second point, standardization.
We need to be able to discuss together. It's like to not be able to talk together. And Internet is a common thing, so reversibility and standardization.
>> MODERATOR: Thank you very much, that was short and sweet, very much to the point. And you actually asked or recommended that we could have more dialogue between the private sector, the multi‑stakeholder community and the Ministers, and we actually do have a Minister also on the panel so that's a good thing. I would like to introduce Minister Johnny Plate, who is Minister of Communication and information in Indonesia. Your portfolio covers many of the main themes of this Forum actually, both the cybercrime, data sovereignty and information technology. We are very pleased to have you with us Minister Johnny Plate.
>> JOHNNY PLATE: Thank you. I was sworn in a month ago and I have to deal with all of this informatics. I'm talking about Indonesia only now. As Indonesia aims to become a digital nation, a great emphasis is being placed on improving connectivity across the Archipelago through infrastructure, fiberoptics, BDS, satellite that connect people from Papua in the east, it took 75 years for telephone to reach 100 million users worldwide, but it took Pokemon Go only less than a month to reach the same number. Private adoptions of universal technology has been a key consequence of unprecedented cross‑border data flows.
Data flows have made our country safer, more efficient and productive. Indonesia enabling of Internet access to its entire populous and leads us to inevitable data governance challenges. As more than 150 million Indonesians now have access to the Internet, Indonesia is operating on a multi‑stakeholder approach to supporting our citizens' data. This approach involves ministries, the national police, telco companies and civil society. I will clarify the roles and responsibility of stakeholders with examples.
First, our Government will soon complete the general data protection regulations, which is currently being discussed in Parliament. This new data protection regulations will not only acknowledge data privacy as a basic right of every citizen. It will also guarantee protection of consumer's data. Second, Indonesia has launched the digital literacy movement in Southeast Asia.
The national movement on disability literacy is a multi‑stakeholder grass roots movement. This movement is made up of businesses, communities, Government entities and academics to engage with and empower communities in data protection, digital literacy, develop curricula and govern cyberpeace. It is taking concerted access against hoaxes, fake news and cyber bulliy that have become rampant.
The two examples illustrate the key roles of the regulator and regulated. On the one hand regulations are given legal weight to the importance of all stakeholders all to give up protecting personal data. On the other hand, citizens must be aware that data privacy is a basic right he or she enjoys.
This can only be achieved through strengthening disability literacy. In my view Government must act to protect citizens at all costs, and that includes data. Civil society must educate itself and be educated in schools on data privacy rights. The public and the private sectors may collect data, but must do so in accordance with the law, and we are just witnessing the early stages of Artificial Intelligence, big data and Internet of Things.
Future innovations will revolutionize daily life as we know it, but only if we follow these principles can we ensure that data protection will be upheld in the wake of new innovations and technologies. We have a long way ahead in answering our long‑term digital competitive and analyzing our data governance framework. Thank you.
>> MODERATOR: Thank you so much Minister, next in line is Theresa Swinehart, the Senior Vice President of multi‑stakeholder strategy and strategic initiatives at ICANN, and you are also a leading advocate of an open and secure Internet and, of course, an expert in global Internet Governance and cooperation.
>> THERESA SWINEHART: Thank you, and thank you very much for having ICANN here, we appreciate the opportunity. We are one of the many players in the Internet space and the Internet Ecosystem and you may ask why we are here on this panel on this topic. It's really about trying to retain an open, stable, secure, interoperable Internet. We deal with the coordination of the unique identifier system, the addressing space, IP protocol, domain name and IP addressing. And that's really the mechanism on which the platforms run, and so this conversation is very important because there is dependencies between frameworks being looked at around data and frameworks that impact the underlying infrastructure of the Internet more generally.
I was struck by one piece of information that was prepared for me in preparing this. It took radio ‑‑ it took television to reach the same market, but for the Internet, it was only four years, and I think that really goes to some of the earlier points about the rapid pace of the evolution, but also the cognizant nature of having all players at the table in discussing frameworks that need to be addressed around data and around the future regulation or best practices for the Internet space.
So if we look at the ecosystem more broadly, we heard some other terms earlier around the importance of the trust in the Internet, mow it works, the security, the stability, and for that you really need to bring all of the players together to have that conversation. There is no single stakeholder, so I think having a panel like this is really quite critical.
From our standpoint it's really informing the discussion that's are occurring, the technical nature of the discussions, the potential for unintended consequences around that, and if I can bring one story closer to home for us has been the recent discussions around the data protection legislation in the GDPR. A well intended legislation and a well intended effort. The unintended consequences around that have been seen in areas that are part of the underlying coordination of the infrastructure space and of the domain name space.
So when we talk about what kind of frameworks or regimes are needed it's really ones that bring all parties together to the table, figure out solutions that are scalable, but also solutions that don't have an unintended consequence for the future of the Internet and benefits that lie ahead. So to put that into context for this panel, thank you.
>> MODERATOR: Thank you also for your brevity. I would now like to introduced to. And you have also been involved in the creation of the relation of the IGF thank you.
>> LEONID TODOROV: Well, I will be short, first of all, I represent a huge community of 64 organisations that run national Country Code Top Level Domains like DE for Germany or FR for France, but across Asia‑Pacific, and as you would imagine the region is huge and so is the diversity within the region because we have huge countries like China or India, and very small ones like, for example, the small Pacific Island of Neva with 1500 residents. With that I must say that we face probably the same challenges of anybody else because we live on data.
Let's say in contrast to Uber, we don't have any physical infrastructure, but some of our members operate name servers, which is hardware, but in effect, we run registries which are lists of people or organisations that register domain names which is understandable. And 70% of our members are Governments which is also interesting because then it means that Governments are coming to the forefront of the problem.
With that I must say that by and large my empirical sense is that across the region, there is a huge lack of awareness of the value of data. For example, if you talk to people in certain countries like China, probably you won't be able to make them appreciate the value of personal data and they would not be on the same page with you. So that means that for some of our members, quite many of them, data collection and data storage and use is not, are not specified in any SLA, service level agreements and are by and large left to navigate the way of their role.
And remember that know GDPR across Asia‑Pacific, so there is not that level of concern about the usage of the data. With that I must say that as any CC, country code level and domain operator, our members face the same processes, she adverse processes because Governments are really keen to take on this subject, and those brilliantly put, brilliantly put by Bertrand Lashappel, three things that really concern us.
First of all, in reacting sometimes irrationally to the challenge, Governments are trying hyperjurisdiction, which means they focus on jurisdictional issues. Secondly, legal plurality, you would imagine how many laws are adopted across the region and worldwide and thirdly, legal arms race. And we are in the epicenter of that legal arms race, although we operate in each and every jurisdiction across the region. So the remedy is simple it seems to me, and that is education and awareness raising. UN put it in one of their documents back in 2003 and I see no other remedy actually. Thank you.
>> MODERATOR: Thank you so much, we have saved the best for last, so I now turn to Monika Wiederhold, Executive Vice President of airline central in Eastern Europe, also care yam of the managing board of am Ma dayous Germany. Monica, do you want to finish us off?
>> MONIKA WIEDERHOLD: Amadeus is the leading travel tech company and half of all of the airline passengers worldwide flow through our system. That means we are dealing with a lot of transactional and private data. What we are trying to reach here collectively is really organising the digital world of the future and that is just amazing task in itself.
So what we see is that we have to organize both sides, the data owner side, and the data user side. The data owner might be a citizen, a person, a traveler might be an organisation, a company, or Government. Both needs actually we think are quite similar. You need to have trust. You want to know what to do with the data, so the owner should be in control of the data. We think that is most important, with it an individual person to privacy reason maybe, be it an organisation more to intellectual property reasons, but the control of the owner of the data is key. On the user side, on the other side, we see that the accessibility and manageability is very important and this is quite difficult and we heard it earlier in the keynotes that even for a smaller, for smaller companies, medium‑sized companies this is an issue.
What I would like to do is to finish with two examples what we do as a company to address both sides, big projects going on on the owner side and the user side. On the owner side, I would say in travel the most burning problem, one of our biggest projects right now is really trying to create interoperable digital identities. This is something in the new digital world of travel which is so important for the travel and it's so difficult to solve because we have all of these different national interests, legal environments, et cetera.
So how can we collectively create and corroborate and secure these digital identities of our citizens, of our travelers in the future? And you really think through, close your eyes and dream of the travel of the future, it will be quite different, you won't cue up at hotel disses or airports.
The travel of the day is really a difficult journey and it will be different tomorrow. And digital identity is key to do so. And on the other side, on the usability side what we do as a company, but what I think a lot of companies do is really open up our data, not the individual lines, of course, but the aggregated ones, the Artificial Intelligence skills we have, open it up and provide an environment for developers to access it. So seen as an individual person, as a startup, you could use this massive data and the massive skills behind to create innovation and businesses on top.
Now, that is to close. Both sides have to be managed and maybe the last word I have is really whatever policies we create, we have to create a practical manageable output for users and owners.
>> MODERATOR: Thank you so much. I think we were due to the lateness of the proceedings this morning, we were given one minute ‑‑ ten minutes extra, I'm sorry, it's only 9 past 1:00. So there is one minute available if any one of panelists thought there was something that they wanted to react to that others had said. Anyone? Perhaps especially I'm looking at Grégoire Kopp, Theresa Swinehart and Carina Röllig who were brief in their comments.
>> Maybe one addition, so as we are spinoff from universities, research was always eye huge topic for us and if I remember the times when our or my research colleagues tried to develop something really, really great new stuff, we are always lacking on data, so we take too much time to get data, to get access to data, and I would say if we can solve this problem, all of this research will be done much more efficient and much better. That's maybe the last I would say to this discussion.
>> MODERATOR: Grégoire Kopp.
>> GRÉGOIRE KOPP: I would like to add that there is possibility to do something very concrete. For example, tomorrow, we will share with the open commission the release of the first code of conduct for infrastructure as a service. Is it is very concrete engagement that we will be able to promote a and share a group in society to get that conduct and it is very concrete. It will be engaged in the European area, and we should work on that because data is very important because if you can't take your data, if you can't move, if you are stuck somewhere, it's finished for you. So it's important to do that too.
>> MODERATOR: With that I would like to thank all of you. It's been a very, very stimulating discussion. I won't try to summarize this, because it's been such a diverse discussion. I put down here digital literacy, awareness, education, standardization, trust, innovation, even freedom, cybersecurity, even cyber bullying, so you were very disciplined in keeping the time, but it's difficult to be disciplined when it comes to containing this issue because it really is so diverse, and maybe that leads to two more procedural conclusions.
One is that I think it shows the importance of having this dialogue as many of you pointed to between the political level, business level, private sector, academia and civil society, this has to be a multi‑stakeholder endeavor if we are to succeed and then perhaps secondly also to move from the conceptual or abstract to the more concise and concrete. That's easier said than done, but I think I felt a common wish that we can move in that direction.
Just 20 seconds on what we do at the OECD. We will continue to our digital work, we provide targeted popsy regulations for companies. We have a going digital toolkit where we try to example best practices when Governments and others confront difficult policy questions. We promise that we have listened today and we will continue to stay engaged. We will continue with phase 2 of our going digital project, and is there a brief comment? I hope it's to praise the OECD because otherwise I won't allow it.
>> It was supposed to be. So I think that OECD is best placed, best positioned to do the work like this because in the past, you created amazing alliances and a great dynamic because people actually, not only people, but Governments were keen to follow the best practices in whatever area we can mention whatever Conventions. And they do so voluntarily in the proper understanding that they are joining a club of leading nations that are in pursuit of prosperity and happiness for the mankind, sorry to say that, but seriously, that's a great job and I think you should just keep it up. Thank you.
>> MODERATOR: Thank you so much. You just improved my talking points to a level that was unmanageable, and I promise that this was improvised and we have not provided fees to any of the panelists. Thank you so much, everyone, I have never encountered such a disciplined group of interesting panelists. Thank you for your insights. I'm sorry for the delays, I think we got a very rich discussion nonetheless thanks to you. Should we give the panelists a round of applause.
(Applause).
(Concluded at 1314).