 Hello, everybody. I'm Pranjal Sharma. Thank you for joining this very, very important session on measuring AI and data in the world of the fourth industrial revolution. We're going to be discussing with very four important stakeholders and talk about how can we ensure that the changes that we see in technology around the world can be measured for a larger good. Let me quickly introduce my panelists. Eric Brinkelson, Jerry Yang, and Akeko Yamazaki Professor, Director of Digital Economy Lab Stanford University, Minister Paola Pisano, Minister of Technological Innovation of Italy, Presidency of the Council of Ministers of Italy, Osrona, Chairman of the Board of Directors, Credit Suisse Group AG, and Urvashi Aneja, Founding Director at Tandem Research. The question that we are going to be looking at during this session is really about what can be done, how does the world benefit from technology, and to understand the benefits you have to be able to measure it. Now, a quick update on how we're going to be running this session. The first half is a public stream that all of you have joined, and in the second half, we will be going into a private discussion for those who have toppling access must stay on. Let me begin by getting right into the discussion, and I'll be updating you on other issues as we go along. Let me begin by requesting Eric for his thoughts. Eric, you've been doing a lot of work and research on measuring it, and you've also come up with a concept of GDPB, which is GDP benefits. In a short brief introductory point, if you can help us understand, what have you concluded with the research that you've done so far? Sure. Well, first off, thanks so much for bringing us all together. This is an incredibly important topic. We're in the middle of a big transition in the economy from goods and services mainly being made out of material things, out of atoms, to them being made out of information or bits. Unfortunately, as you suggested, Rangelo, the current measures, including GDP, are not very well designed for this world. Going back to Adam Smith, a lot of our economic principles are designed around a different economy and GDP, which was one of the great inventions of the 20th century in the 1930s by Simon Kuznetz and others has many strengths, but it's particularly bad at measuring the information economy. Let me give you an example of what I mean by that. If something has a zero price, it has essentially a zero weight in GDP, because GDP measures all the things that are bought and sold. But a lot of the value we get, we're now here on Zoom. A lot of us use Wikipedia or Search or all these apps that you can get on a phone. They have zero cost, but they still create a great deal of value. What we've done, as you noted, is create a new metric with my team at MIT, and here at the Stanford Digital Economy Lab, that we call GDPB, where the B stands for benefits. The idea is instead of measuring what it costs to buy something, and therefore it's contribution to GDP, and by the way, productivity is based on GDP, traditional productivity, we seek to measure the benefits. How much a consumer would be willing to pay if they had to to get that good? For instance, how much would I have to compensate you to stop having access to Zoom or start having access to Wikipedia or the different apps? That value is a measure of how important it is to any given consumer. If you sum up those values across all the different goods, both digital goods and non-digital goods, you get a measure of the size of these contributions to economic welfare. What we've found is that the digital economy is not surprisingly worth hundreds of billions, even trillions of dollars. If we have a better set of metrics, we're going to be better able to steer policy, to steer decision-making by CEOs, and understand better where the real source of the value and innovation are in our economy. I think that's a great point to move to Minister Visano. Policymakers, it's very important to minister to understand the impact of technology. Technology is not just about enterprise as we've seen, but governments across the world are now deploying technology for various civil society services, for benefit, welfare schemes, and also to measure the impact. How do you see and what are the efforts that are happening in the government in your example, for instance, to ensure that policymakers also have the right framework to be able to understand the benefits or the impact? Hi, hello, everybody, and thank you very much for the opportunity to speak here today. I think that some challenges may be achievable only through international cooperation. Multilateral discussion and initiative can serve as a turning point in achieving common standard. International cooperation can enable it only by an ounce to common value and mutually beneficial, although the advantage of cooperation is not always easy to reach a wider consensus. As government, we should create a framework of trust and knowledge so as to avoid crystallizing unnecessary resistance or fears around digital technologies. For instance, together with FAO, Microsoft, and IBM, we signed the wrong call for AI ethics that is an initiative promoted by the Pontifical Academy for Life, and this document contains a set of policy principles such as bias mitigation, transparency, inclusiveness, security, and privacy to guide the development of artificial intelligence algorithms. And we wish to take the opportunity of Davos' dialogue to invite other countries to join this call. Moreover, we are among the funding members of the global partnership of an artificial intelligence that JPEI, an emerging forum for international collaboration on AI, which is informed by the OECD AI policy principle, and promotes cross-pollination of OECD's policy discussion with experts from industry, civil society, academia, and governments. GDPI's experts are collaborating across five working groups. Teams, the first one is responsible AI, the second is AI and pandemics, the third is data governance, the fourth is the future of work, and finally innovation and commercialization. One last note, cooperation among countries require the engagement of citizens. Individual should be aware of where AI is used and applied, and if the technology is just for the elite, we think that it's a failure. So, towards this end, we introduce an online course, Elements of AI, an open source course, created by Finland, an adapter to the Italian language, aiming to provide the basic training elements for everyone who wants to deepen their knowledge of AI technologies. And to conclude, I think that international cooperation has its benefits, but it's a longer-term endeavor. Mr. Besaro, those are very, very relevant points, and I think it's very interesting that while we're discussing technology, words which are coming across our words like collaboration, citizens, governance, and I think that's what's going to be defining technology in the era of fourth industrial revolution. I'm going to come to you, Urs, about the role industry and business leaders have to play. And now the amount of data which is being generated across the world, almost every human activity, every transaction can now be captured. So, we are in some ways going to be flooded by data. The question then is how do you analyze it? Do we have the right structures? Do we train the right people on doing it? And again, in different sectors, the kind of data that emerges are very different. So, we need common standards and to be able to understand it in a way that there is a harmony in it. How do you see this evolving words, especially with your experience? Thanks, Pranjal. I think it's a very good question and I think the starting point would be that, you know, is really the question of the different types or the differences between types of data from distinct sectors and in varied contexts. And the answer would be yes, there was differences, I would say. I think you're getting a little bit lazy about how you refer to data, often describing it as fungible. It really isn't. Data is an ecosystem of subtypes. And this is not just a question of, I would say structured versus unstructured data. There are these differences in volume, sensitivity, entropy, in other words, how long a data remains relevant or not, and value to name just a few categories. Give you an example from my industry, from the financial service or banking industry. In banking, for example, we deal with a range of sensitive financial information that identifies our clients. We have to treat this with great care because our clients and our regulators expect this of us. I mean, it may even be a legal obligation to do so. I'm the chairman of a Swiss bank. Can you expect me to take this very seriously? And I do, so does my bank. And yet not even this data is a simple matter. Some is only sensitive in context. Give you an example, in other words, the data may be separated out in such a way that it makes it much more usable without intruding on our clients' privacy in any way. And there are other ways of sanitizing or anonymizing data to make them available for data analytical work that could never be carried out on the source information. And from a regulatory point of view, sometimes we are compelled to provide this kind of data without being able to identify, for instance, the client behind it, for purposes of regulatory oversight. So the same kind of data can have completely different shapes and forms and be used for different purposes. And the field of privacy-preserving data analytics is one we ought to all be paying attention to. It allows us to both protect the privacy of data that requires it as well as to make proper use of it and build our businesses to give just one example. And for another, I would expect this soon to be able to give very strong guarantees about the privacy of individual medical records and yet to make, on the other hand, extensive use of the information in those records to promote better healthcare for all of us. So I think that's something which we have to grapple with, which on the one hand I'm sure governments will have to deal with it. On the other hand, I think if you look at research in the pharmaceutical field, for instance, it's absolutely crucial that you get to as much data as you possibly can in order to promote better healthcare. So those are the challenges I see, I think also from an industry point of view, and it differs vastly. So collaboration is not just between industry and government, but also it's within industry because now the walls between sectors are also collapsing. So financial technologies basically means both telecom and finance are together. So how do you make sure that the old structures are changed and new ones are put? But this also brings me to the other point about the gap between perhaps at least the perceived gap between developed and developing economies. What I think is fascinating about the fourth industrial revolution is that the technologies are far more egalitarian. You don't have to get into issues of technology transfer. So the developed world doesn't have to sell to the developing world, but everybody can solve for themselves to a large extent using data and AI and such technologies. Urvachi, in your point of view, if you look at civil society, especially in emerging markets, how can this measurement make life better for policy makers? How do you ensure that can we perhaps then target welfare activities in a better way? Or do you see that the challenges in developing markets, emerging markets will be distinct from the others? Anjali, thanks to the organizers for inviting me here. It's great to have the opportunity to discuss this really important issue with all of you. Let me start by maybe pointing out something that one of my historian friends often reminds me that the fourth industrial revolution is probably the only revolution that has named itself before it even really started, or that has named itself in its infancy. Most revolutions are named in hindsight, not at their beginning. So I think, and that's obviously, I mean, at likely, but I think that also points to something, right? That to some extent, the fourth industrial revolution is also a vision. It's an agenda, it's a narrative. So we need to ask who is promoting it, who benefits from it, and who are the losers of this narrative, right? And revolutions, if we think of this idea of revolution, typically reorganized political, social, economic relations, right? They redistribute power. And the way I see it, the fourth industrial revolution in its current form is not resulting in a redistribution of power. In fact, the growth of the AI industry, as we see it today, is predicated on the concentration of power with a select few actors, and is in fact accentuating that concentration of power. So to me, to go back to the metrics question, that's a really important metric that we need to think about. So alongside GDP, we also need to think about the distribution of technology gains, the distribution of power in the context of the fourth industrial revolution. That's particularly important from a developing country context. So I think in many countries in Europe, particularly, we see risk-based approaches to regulating AI, right? And I think risk-based approaches work well for fueling innovation. But if you look at what's happening in the global South, AI and other emerging digital technologies are not just fueling innovation, they're a critical part of the development story of these countries, right? They're supposed to be enabling developing countries to navigate complex socioeconomic challenges. So if they're part of the development story, then I think a risk-based approach is gonna fall short in being able to achieve some of those development outcomes. So with that kind of, with that in mind, I think some of the other metrics that we need to introduce into this conversation from a developing country perspective, particularly, one, it was the distribution of gains that I already mentioned. I think the second is around the impact on individual rights. Privacy here I think is a really good example. We don't have time in this conversation, obviously, but I would love to know another point on what privacy-protecting data for an example is and how we get there, but because the current kind of frameworks around anonymization and consent don't quite seem to be fit for purpose. I think we also need to think about structural impacts and harms. So for an example, what is the impact from marginalized communities to be able to access opportunities? How does state-society relationship change? What does it mean for the accountability of the state? I think we also need to think about the environmental impact. That's something that doesn't get discussed enough in some of these conversations. So training a sophisticated AI model for an example, apparently pumps in five times more carbon dioxide into the atmosphere than an automobile in its entire lifespan, including manufacturing, right? So we have to include these additional metrics if we're not thinking about AI, particularly in a developing country context, because it's not just innovation that's at stake, but it's the development trajectories of these countries that are at stake. That's a great point, Roshan. And I'm going to take this to Eric. Eric, in one of your papers, you also very rightly said that there are optimists on one side and pessimists on the other. And the optimists are the investors and the technologists and the pessimists are possibly the policymakers and the civil society. And you see a little bit of the reflection on what Roshan just said. I think getting some kind of a harmony in the way technology is measured, is missing. And that's what we heard from us and we heard from Minister Pisano as well. Is there, do we have to create a completely new model for this in terms of doing it at a global multilateral level or does each country have to go and create its own? Well, I want to underscore what Orvashi said. I think, and what you're saying, there are optimists, there are pessimists, but I think that both of them make the same mistake, which is that they often think that something's just going to happen and that the rest of the world or the technology is going to either turn out good or badly. Really the core is our agency. And I don't know that it's so much a matter of a new model as being in a line with our values. We have more powerful tools now than we've ever had before, which means if we want to, we can shape the world more than we did in the past. And frankly, we haven't been choosing to shape it in a way that leads to more equitable distribution for most of the past 20 to 30 years. The economic pie has been getting bigger. Technology is creating enormous amounts of wealth, more millionaires and billionaires than ever in history, but there's no economic law that says that that's going to benefit everybody or even benefit most people. I want to be clear that for most of history, it has been broadly beneficial, but for the past 20 years, a lot of people have been left behind. And I don't think that, I don't blame that on the technology or any particular model. I don't blame that on some choices that were made by governments, on tax policy and other things, some choices that were made by CEOs and a lot of other people. And if we choose to, we can actually have a much more equitable distribution. I think, as Rubashi said, the digital technologies, some people talk about the fourth industrial revolution have the potential to have much more widely shared benefits. Digital goods have some three characteristics that are not common to goods made of atoms, as I was describing earlier, they're free, perfect and instant. They can be made at zero marginal cost. Each copy is an exact identical replica, perfect replica of the others. They can be instantly transmitted anywhere on the globe. So most of these digital goods could be available to everybody at almost zero cost broadly. It doesn't necessarily go, automatically happen that way. In many cases, these goods have led to more winter take, all markets, more concentration of wealth and power. But those are some design choices that we have, with that many people have made and policies and intellectual property and elsewhere that we've made. But it's certainly not the only way to organize things. So I'm glad that the World Economic Forum is bringing attention to this. And I think hopefully many of the people listening to this will think a little bit about not just how we create more wealth, but also what kind of, how we want to use these amazing digital technologies to create broadly shared prosperity. Let me go back once again. I said, if you look at the access, I can tell you for India alone, we have about 750 million people who are now connected to the internet. So we call it digital inclusion. And a lot of them, at least 30 to 40% are first time users. So for them, the concept of privacy and data et cetera don't exist because they're just so happy to be part of the mainstream that they don't even realize what they're giving away to receive what they get. But this is why, Minister Pisano, I want to come to you. We've had a very interesting report on global governance and technology recently. And while there are divides within countries, between countries, the question then is, what kind of global governance can come? Europe has done a great thing by setting the benchmark with GDPR, but technology companies who have a lot of power are saying that, well, we'll behave differently in Europe, but we'll behave differently in Africa or in Asia or in the Americas. How do we get the world together on a common global governance framework? Yes. I think that government should ensure that data governance and decision-making processes follow a sound approach by assessing constraints, risk and rules, and surrounding data sharing, collection and use. We can play a major role to ensure criteria such as fairness, safety, robustness, respect for human rights. We recently launched the National Digital Data Platform that is a single platform to which individual administrations can communicate and share data and APIs in a free and easy and open way. That's allowing the creation of new services and data application tailored on the needs of citizens while fully compliant to privacy and security regulations. The data platform offers a common access to data and makes the transfer of information and data on different parts of government more efficient, transparent and less costly in order to achieve public sector productivity gains and more effective service delivery. Intelligent data collection and data analysis have been crucial during the present COVID-19 epidemic which has been at the same time an accelerator for digital technologies but a disruptor for our way of life. We experienced major challenges, vertical silos in public administrations made this cooperation among levels of government quite difficult and coordination and data sharing is not an easy task even within the same country. So the international coordination of these issues is even more complex. I think that data flow is the lifeblood of the digital economy. It is a key requirement of the development of AI solutions but both trust and privacy are the key component. The discussion is not an easy one and it has been so for a few years. And I think that there is a legitimate reluctance to share data but also an unwarranted tendency to restrict data flows internationally. To help overcome ethical and legal barriers international organization like WF or OECD are important partners. And the WG on data governance of the JPI Global Partnership of AI has developed the interesting idea on includes recommendations to announce international cooperation on data and providing a more project-based direction. These conversations can all help governments to get a better understanding of the reasons for which the barriers exist so that it will be easy to converge. Right, thank you minister. We have about five or six minutes left. So I'd like to request both the earth and we'll worship for your quick statements in the next few minutes left. Earth, the question is now also about power of concentration of power. There was a while when everybody thought that the old power rested in the global banking and financial companies what looks like you have competition. And now all the power has shifted to the technology giants of the world in terms of being the most influential. What is the responsibility of the industry and business leaders to ensure that we are able to use data AI in a more responsible way for a logical. Obviously that's a very big question which obviously is difficult to answer in just one minute. Obviously I think that the issues or the priorities for let's say for the industry shouldn't be vastly different from what governments have said. I mean we have an interest that data flows are harmonized for obvious reasons because it creates a lot of value if it happens and it creates a lot of these synergies or a lot of costs if it doesn't. I mean, give you one example of the ongoing discussions between United Kingdom and the European Union about data adequacy alone would cost roughly over a billion a year just in terms of additional costs that they just totally unnecessary. I think the other thing is we have to be cognizant of the fact that data and flows may post challenges that have to be overcome. I give you one example. For instance, on the GDPR individuals have a right to be forgotten. And on the other hand, there are often anti-money laundering obligations on banks for instance to keep exactly the same data for good because you want to be able to make sure that you can fulfill your obligations there. And if I can use that angle a bit for the industries and this is true for the large technology firms as much as it is for banks or financial services companies I think we have to see that we have sort of a common level playing field in terms of what we do with regard to data protection, data security as well. And then the rest I think will be done by the industries themselves. They have to change. Obviously banks have vastly changed in the last 10 years and will continue to change over the next five to 10 years significantly in terms of how they look, what they do, how they do it, why? Because they got disrupted. They got, let's say, the challenge from technology firms be it smaller or larger technology firms. And this is a trend that will continue in the end. I think that's helpful for society at large but we need to be sure that we address the issues. We address the issues of how we run data but also how we protect data. And that's something which I think is still an answer to a large extent. I wish you the point about preparing civil society for this is also important. While there's a responsibility on the government and the private sector, I think there's a crying need for awareness on the importance of data and value did talk about the concerns. Is there now a crying need for us to ensure that those who are joining the digital mainstream across the world should be made aware of what their responsibilities are and what the risks are? Yeah, I mean, I think just very quickly to totally echo what Eric was saying, right? If we do have agency and public policy has a really important role to play in steering technology trajectories to align with societal goals. I think the civil society capacity point is really, really important. So if you look at some of the conversations around AI, one of the things that we ask of in terms of responsible AI is that it should be transparent, right? That's one of the values that we hold here. But the flip side of transparency is actually expertise. You need to have someone who has the expertise to be able to evaluate that transparency in a meaningful way. And if you don't have that expertise in civil society, that transparency will not achieve what you hope for it to achieve, right? So ensuring that we're investing in civil society to exercise and influence that agency that Eric is talking about is really, really important.