 Welcome to ITU Telecom World 2018 here in Durban, South Africa, where I'm very pleased to be joined in the studio today by Jacques Bugin, who is a senior partner for McKinsey & Company. Jacques, welcome to the studio. Thank you. Now, I'd like to start off by asking you a little bit about the fact that you recently worked together with ITU for research on assessing the economic impact of AI. Perhaps we should talk a little bit about what new technologies mainly promise. Yes, in fact, we're working with the ITU on a large set of technology questions. And obviously what people are interested in is obviously technology that's going to be disruptive, not basic technology that does not do anything. We look at big and disruptive. Now, the one we are looking much more into detail has been AI. AI is obviously a very generic term, but it encompass everything that you can do from a digital technology point of view. It includes technologies like facial recognition, everything which is NLP, everything which is robotics, physical robotics, with those of smartness into it, driven, of course, by the fact that we have so much of big data. These big data feed algorithms and these algorithms start to know linguistic type of technologies. They start to learn by themselves. So we look really at AI as the focus of this. Why? Because again, they are really dependent on the networks and obviously the networks is the bread of butter of ITU members. And at the same time, because they start to be extremely efficient and they start to scale. So it's the time for AI to go beyond connectivity. Talks about ITU members. What's the benefit of being a member of ITU? Well, I'm not a member of ITU. I'm cooperating. I think from a perspective of what the ITU does, they do obviously a lot on the technology innovation part and the standardization. I imagine that in the context of AI, everything's going to be not necessarily standardized, but at least synchronized. And I imagine that in AI, given that AI will actually develop and deploy to many sectors, i.e. car manufacturing, there will be a lot of standardization. Not necessarily standardization on the technology side only, but standardizations like if a red car, which is fully automated, drives over a green car, which one has priority? All these kind of new questions coming from the fact that we need to define new rules to robotics, to artificial intelligence, all these rules need to be defined, agree upon, and maybe the ITU can play a great role on that. Okay. Let's talk about emerging technologies. How can they be leveraged, do you think, to meet our needs in terms of sustainable development? And what would you think will be the economic impact of AI in our society? Yes. In the report we have done with the ITU, first we focused especially on a subset of the SDG. As you remember, SDG has 17 objectives, which I think is quite comprehensive and quite versatile, but at least we look at the one where we have something to say, which is pretty much the economic impact. And that means that in the economic impacts we look at the ability to grow, ability to expand economies, but in the same time we look at some of the SDG like inequality. All this technology is going to create welfare for everyone, only for some, and at the expense of others, these are fundamental questions. So we looked at that. In the report I think we made six new findings. And again, there are a lot of experts talking about AI. Some people just draw figures at us, or at you, and say it's going to be big, but all big really. And so what we've done is that we have tried to answer a few questions with, again, six inside. There will be an insight about all big or fast. What exactly drives that? Is it automation and you have no job? Or is it innovation and you have possibly new kind of jobs? We need to understand the mix. There is a question, as I said, about who actually, is it the same for everyone? If you happen to be US-China and you lead the war for AI today or the contest for AI, is it going to be different if you happen to be Brazil, for instance? That's the fourth question that we look and get some insight. There's a question about, we talk a lot about technology race. Is it a technology race, a competitive race? And last but not least, there is a question, as I said about the SDG question, who's going to benefit from that? Now, let me give you two or three points on each of them. How big is it? Well, based on the average of the thousand of scenarios we run, we come to the conclusion that this is 13 trillion of equivalent value of economic activity, monetized economic activity upside, that pretty much 20% more than what it is today. Now, from 2017 to 2013, if you are good in maths, that means 1.2, 1.3%, a year of additional growth. Now, if you happen to be European, like I am, 1.2% of growth per year, it's very welcome these days, especially because we have an aging population and at the end of the day, we have seen a productivity puzzle, productivity starting to actually decline and not to be sustained. So, 1.2, I take it. And by the way, 1.2, how does that compare to the past? We have come to multiple of those revolutions, the first one being the steam engine, the second one being electricity, give and take, we think it's comparable, if not larger than what we have seen from these technologies. And again, it's percentage. A percentage at that time was much lower than percentage we have today. So, we believe that AI is a set of fundamental technologies from which, by the way, ITU and their members are going to need to actually activate the supply chain element. Sizable. Second question, by when? Here is the point. AI is already among us. Possibly, you don't know it because, again, like electricity, you don't touch it. Electricity can hurt. But on AI, look at search. Most of what you do on search, whatever providers you use, most of the time, it's actually optimized to AI. Not only, by the way, for finding the right answers, but to make sure that this answer is fastest as possible. So, the user experience is, wow, I ask, I get. Do you like when you ask and it takes five minutes? No. So, you need both. So, AI is already there. So, from the timing point of view, there is a lot of use cases, which, by the way, help us to refine the numbers, I tell you, because we got evidence of the use cases. And the timing is such that it will come quite fast. No. Again, here is the catch-up, the catch-22. The catch-22 is that if I want to give you 100, of course, if you have it right away, you take it. But usually, I want you, at least, to make an effort to get it. So, that means you have cost of transition. There is no way you can do, you buy robots and robots will do the work for you. Robots need to be designed in such a way that they work in your workflow and your shop floor. It could be that you need to change your IT architecture, because the cloud and the data are not necessarily the same that you can get from Internet of Things and so on and so on. And, by the way, there are externalities, which alludes to my last point of inequality. Some people will not be fit to work with robots. Some people will not have the cognitive capabilities to actually work with software of certain type of artificial intelligence tools. And, by the way, there will be new jobs, as I said. We possibly need multiple professors of not psychology and philosophy, but AI ethics. That is cost. Now, if we put the two together, here's a striking fact. You will not see on the global scale a high visibility of number of AI. And you will come back to me five years time and say, Jack, where are these numbers? Are you telling me? Well, these numbers are actually in the economy, but there is cost and there is the upside. And, in fact, all those scenarios tell us it will take five years to see this underground of technology to become so powerful. That means should you care? Yes, because, as I said, it's today it's happening. If you don't do it, you will not be part of the 80, or the value that we have said, the 13 billion that we have said. So, the timing is today, but it will take five years to see who wins or not. Comes the third point that I say it. The third point that I say is that who's going to benefit? In fact, everybody can benefit. The technologies can be hold out and adopt, but some countries have faster than some others. US and China is definitely ahead of the game. Europe is in the middle, and like usual, you see a bit of a crescendo in terms of who is first and the like. Why? Well, first, they rely on digital technologies, and digital technologies are not yet fully rolled out, let's say, in Africa or in part of Asia and the like. So, this is obviously a catch-up phase that needs to happen. As I said, it requires to get very strong firms that ingest these technologies, and obviously, the largest multinationals tend to be in the US and developed countries. You need skills. Do we have all the skills that we need? So, the point is that this technology will converge from a performance point of view, from an adoption point of view, but it takes time. And if you don't upgrade the rest of your ecosystem, you are likely to be with the AI divide as much as digital divide. So, I pass directly to the last one, who wins, who loses. I told you, the guys who wins is actually everybody potentially, but you need to upgrade not only the technology, but the ecosystem that works with that. Second, that means what? Some companies will be faster and some others to adopt. Why? Because they are foresight, they know how to master the capabilities behind that, and those guys will actually do two things. They will be innovative and possibly they will take market share from the one that is not very good in using AI to be competitive. So, there will be winners and losers in firms. We estimate that if you are a winner, you can grow five points per year better than anybody else. In terms of economics, these are what economists call the superstar firms, having 5% every year attributed to technology for the last 15 years. I'm quite sure if you were private companies and you will be my shareholders, you will sign if I give you that plan. Now, the worst thing is that some companies will likely not adopt these technologies because they don't have the digital skills, they don't have the infrastructure, they tend to be more small medium enterprise, they tend to be in the service industry simply because the scale is different, and these companies will suffer, some of which will possibly get under the path of their own profitability. So, it's a watch out, it's a watch out. And the last but not least, what about us? All the discussion we have on AI is always about what? Employment. Should we still have employment? Now, I don't know the answer because that's a future, but I can tell you one thing. Every technology is always skilled jobs, so this is not a set of new technologies that will definitely kill job. It's actually the reason it's not that they want to kill job, it's simply because if you ingest them in companies, most of the time it's because they can do work slightly better than some of the times that you do. But in exchange of that, they will create new products, new services. Again, search is a new product and I bet you use it a lot, you know. The mobile is used by so many people and the first thing we don't want to lose is our mobile, much more than our wallet these days. That's quite fascinating as a story, which is people prefer to lose their wallet than their mobile, where obviously the mobile is much more personal. So, the point is that this innovation will and the net net will be quite nice. So, the point is not employment. In fact, I go back to my first point. We will see 13 trillion, possibly with the same level of employment as of today. Now, between today and 2030, in developed countries, we don't care too much. We are aging and we don't create babies. In developing countries, the challenge is that you still grow that force. So, yes, they're possibly going to be part of the AI divide and pressure on that one. And the last but not least, if it's not employment, the issue really, what is the issue? The issue is skills. You know, this capital being artificial, becoming smart, one day replace what you do that's not necessarily fully that smart. And at the end of the day, that means you're going to get a job and task to work, you're going to need to upgrade your skills. And the final question here is really a scalability issue, which go back to your first question, technology does not rule the world. Technology actually obliges us to be better because the better the technology is, the better our scalability needs to be. And there obviously the challenge is quite important because today, believe it or not, even in the US, 6% of the people work in creative industry, but robots cannot be creative. It can replicate, cannot be creative, but only 6%. And 50% of all time, we spend on repetitive tasks, we spend on manual tasks, or in basic cognitive skills. That means that we still have a long way to go to exploit or smartness, and in fact, the tools of technology can enhance that smartness much more than anything else. And that's the future. So those insight is pretty much what we have in the report with a great way of saying it is that a set of technology we are going to use, not to abuse, but if you use it for the good, there is 13 trillion after all the cost taken, possibly with some challenges of who receive it, but with right redistribution and the like, I think we go for a good AI and a better world. It certainly sounds very positive, and we will better step up to the mark. I think we're going to take advantage of this. Well, thank you very much for that. Just to finally, just what's the value of attending events such as ITU, Telecom, World for You, I mean, there's lots of places I'm sure you could be at. First of all, I would say it's my first time obviously at the ITU, and I was quite surprised by how vibrant the place is. You see so many types of different person. We got so much value of exchange between a government member who is a Telecom executive and the like. And fundamentally, it's a great place as well to see all vibrant part of the economy, of the global economy, especially when you come here. We got all African economies represented. We got great discussions about how easily they use these technologies, and it's really show two strengths of these events. The very first one is that you see people that otherwise you would never see, and the value of exchange is great. And two, I learned a lot about how quickly Africa is using technologies and leapfrogging on the Telecom side to basically create a new market of Telecom. It's quite eliminating. Thank you very much indeed. Thank you. Thank you.