 Field philosophy has a bad name and I'm not sure about this group, but in general if you walk around and you say you are a philosopher then you look at you suspiciously. Because what do you really do? And normally the answer is that you teach philosophies, okay that's fine. But if you do philosophy as in doing philosophy, in除 teaching it, you really are suspicious. The reason being that unfortunately philosophy goes ups and downs. The history of philosophy if any of you has read anything about it or has deux is ben exposed and I may be actually talking to colleagues I don't know and I'd rather not know it goes into waves, it's not a history that has a very nice straight line of developments and contributions but it's more like great ideas which then become increasingly scholastic and self-serving, it becomes a discourse. I talk about your paper, you talk about my paper and it's that little problem we move philosoficiol os ydych chi'n philosoforiaethau, rhai gynhyrch yn argynnu'r wath. Rydyn ni'n hyn i gyd o'r dysgu, o'r rhain syniadau syniadau, y rhai gyfafloedd dros y gallu cyfarwod hwn yng nghymru â'r problemau philosofiol. Ond rydyn ni'n gofio ar gyhoedd, a'r ddech chi'n gwybod i'r ddaf. A'r ddaf yn ddwy'r ddod o'u ddysgu'r ddechrau. So, diluted, it becomes internal discourse, and then up again. So, until recently, we were into a sort of scholastic ivory tower stage, where philosophers, even these days, in fact, if you see what we're publishing, specialized journals, are utterly irrelevant. We made ourselves irrelevant. We told the world that all philosophy had to do was to clean up the language and possibly show that some problems that everybody feels are important are in fact misconceptions. So, solving problems by eliminating all the solving problems that was the agenda. Once you've done that, you clearly are redundant. If there were any coherence, we should be closing philosophy departments. Fortunately, that was not the case, as in that was not the task of philosophy, and therefore we're not closing philosophy departments either. So, what I'm going to try to show you today is that there's plenty of philosophical work that needs to be done in our society, information society, and at this particular stage of development of human culture. Any philosophical discourse that tries to retrench behind some kind of ivory tower is not just pointless, which wouldn't be that much harmful. I mean, there's plenty of pointless research all over the place, but it will be irresponsible, and then the judgment is harsher. As in philosophy, we'll be living, and philosophy, I mean anything not necessarily done by philosophers professionally, and if critical thinking about open problems, we'll be leaving the ground empty to other people with worst messages. Fundamentalist, gurus of all kinds, and as with philosophy, has a responsibility not to retrench behind the ivory tower. So, speaking of ivory tower, there's plenty of ivory tower back in the place, but the kind of work that we're trying to do at the Oxford Institute is a little bit more engaged with the world and the information revolution. And that's what I want to talk about today with you. The usual line is that you don't have to agree with me, but if we disagree, we better know about what we disagree. So, disagreement is welcome as long as it's not based on misunderstanding. So, in terms of avoiding misunderstanding, I'll tell you very quickly, because I know you know, so no patronising on my side, just to be on the same page, a little bit about the information revolution, and then two items. One, time has changed as in not the physics time, but time has been understood by us, our age and space. In other words, our environment. So, our age and our environment, how they have changed, and then how that, basically one, two and three, have affected our self-perception, how we understand human nature. With that leading two, and I hope it will be a matter of Q&A and discussion, some challenges that the information revolution, a reconceptualisation of our environment and our age and our nature at this time and in this environment, have led to. Going straight to something that you know, so just a reminder in case the coffee was not strong enough. This is iconic, is Wikipedia level, everybody knows. That's the processing power and how it has developed more slow, etc. We're not quite sure whether it will keep going this way, because physics has some limits, unfortunately, but to the best of our understanding, this is going to go for another while. What is less visible is the other side of the coin, how little everything costs, how much all their power has become so cheap. As cheap as this, and I'll make sure I don't fall all over the place, suppose you have an archaeological piece of technology in your house, an iPad 2, not 2010, old stuff, that thing was able to process, and do we? So you're around what, in the middle? In the middle. It's not very strong. So, finger will do. An iPad 2 had an average power in terms of processing power, 1,600 million of instructions per second, MIPS, which by the time you finish that sentence, it's a lot of processing power. So every second that goes, it had 1,600 million of instructions already processed. Great. Make that equal to $100. That's, suppose that that's what we cost to buy this particular power in 2010. Well, this is down here, 2010, and that's the cost on iPad 2, processing power. This is the cost of that same power back in the years, 80s, 70s, 60s, all the way down to the 40s when we started all this wonderful revolution. The very top, sorry, the very top number here, that 100 with all those zeros, the joke is that only American generals can read those numbers. That's dollars that you've never seen in your life. I mean, it's not HSBC kind of dollars, it's just beyond belief. That in fact is higher than the equivalent in 2010 of the GDP or QVAT. So you start getting a sense of what you're handing your hands in 2010. We should have been way more respectful of that little gadget. Where all that power went into generating trillions of billions of fantastic gazillions of data. And by, I mean data, I don't mean information, meaning that the conversation on Skype, those little cats on Facebook, there's all data, data, data, stuff. For those of you too far away down there, I apologize, but we were all born outside this particular circle. The internal circle is 2009, let's see, over here. If you read, if you see the red dot, 2009, 0.8 zetabyte of data. Since the day we started scratching cows on a cave, that's from that time until 2008. 2008, 2020, 35 zetabyte. That's immense. That's 35 times more than we ever generated in the past. Now suppose that that's wrong by a large margin, 50%. That's still staggering. That's still amazing. Suppose it's only, no, it's not 35,000, not 35 times. It's only double, okay, that's amazing. I mean, we produced more data in the last decade than in the whole human history for millennia. Of course we don't know how much of that is total rubbish. As I said, advertisement, half the money you spent is badly spent. You just don't know which half. The other problem which we've not discussed today is that, and I haven't seen discuss widely, but we should, is that all that grey half circle there comes with a sort of baby boom retirement sort of label. We created all the data we have basically in a decade or two. That's all going to grow up together, get old together, and become unreadable together because of support, failures, et cetera. That didn't happen to books which thousands of years to accumulate in manuscripts, but all the data we have, the digital ones, they all have the same more or less timestamp. That is going to go together. There's a whole army that's marching in line. It's going to cause some problems, but we'll see this in the future. So there are no limits that we can think of in terms of how this stuff is going to grow, apart from thermodynamics, physics is getting a saying about how much you can do with all this, intelligence hours, the only one available in the universe as far as I'm concerned, and memory, memory as in support, the stuff where you put the data. This is the data that don't float in the clouds despite the phrase. Of course all this, and I'm just preparing the ground, so I know you know, so bear with me, comes with lots of problems. Acquisition and storage, where you put the data. Usability, once they are there, can I use them when, how quickly? Security and safety, are you sure that I am the only one who can actually use this data so well? Accessibility as in, well, suppose that they are there, but it's very hard to get them. Accessibility being also a huge right to be forgotten issue, as opposed to availability, analytics, for those of you who actually deal with this black magic art, all this constrained by law and ethics, and of course at the bottom of all this, the usual variable which affects everything in life, money, because anything you do here will come with a price. A price in feasibility, a price in cost, a price in new machines, a price in hiring people, a price in not getting the right thing, opportunity cost, et cetera. So this is the picture, and just to again give you a sense of, we could go in many different directions, this is one we will not pursue. This is all data, very old, but it's the only data I could find about the difference between the data, forget about information, the economist being a little bit generous with the word, the data we are generating, remember those 35 gazillions of things, and how much support we are producing in the world. The essence of my phone is full, I cannot put more pictures in it, so I better erase something if I want to take another picture, well that's a planetary issue. Since 2007, we haven't produced enough support for the data we are generating, meaning there's not enough memory where to locate the data. With a simple result, particularly, that on a global scale, not you and me can afford more, no I just buy another USB, on a global scale, either we do not register something because there's no space in the first place, we don't produce it, it can't be put anywhere, or is first in, first out, those nice emails from two years ago, who needs them, they'll, boom, go, never exist in the first place, or there's a struggle about who gets what amount of memory, and if you are in any department where the ICT office says, well, well, well, this project cannot go there because the hard disk is full, so oh, but it's my hard disk, why don't we kick out that old project, which we don't need, so there's a struggle there that can be manipulated. But that was just as a way of interaction, that's where we've been for the past few years, has generated huge changes, one in time and one in space, which I want to share with you. The one in time is, again, you find all this in the book, so it's a bit of a self-activitise. Here's a suggestion, which I hope is sufficiently reasonable. So textbook material, not my own, if you disagree with me, you disagree with, so the textbook, which is fine, but just in case. Prehistory is defined as any stage in human development when there is no way of recording the present for future consumption, there's only oral culture. Grandma said so, and that's why I have the recipe. But if I forget, that's it, the recipe disappears, nobody knows, because there's no written recipe for that particular thing. So it's prehistory and history and hyperhistory, they work like adverbs, is how you live, not at time in some magic sort of scale. There are still some very few people probably in the Amazonian environment, which a few tribes that live in prehistorically, they have no way of writing basically. 6,000 years ago, here in China, we developed writing and we moved from prehistory to history. Now history is a stage where individual and social well-being, you and your society, they start getting better also because there is some way of communicating, a way of, no, that problem has been solved by grandpa and it's written here, so I can do it again. Or no, no, no, no, cheese is somewhere in marble and then you get somewhere, et cetera. So individual and social well-being starts being related to ICT, but it's not yet dependent on ICT. Now by dependent, I don't mean, oh, of course we need food and shelter and, of course. But today in a society like this one and some other Europeans ones and some others in North America and so on, the development, what makes it the added difference is exactly ICT. Now if you think that that's too philosophical, here's the counter sort of balance to the speculation. Any country that can be subject to cyber war is a country that depends hugely on its ICT infrastructure. If you can be harmed by a cyber attack, clearly you live in a hyper historical society. A society which depends, which individual social well-being, on things of air space or hospitals, databases and so on. So in a hyper historical context, those who live by the digit, they die by the digit just to not be a memorable phrase from another book. You know that we're moving into a different stage because there are plenty of attempts, failed attempts, as far as I could say, to make sense of our culture. You wouldn't have a cyber culture movement, a post human movement, a singularity rubbish if you didn't feel that, oh my goodness, something is really changing here and how do we make sense of all this? I did say on record singularity rubbish, just in case. As I will show you later, I mean just not bad, bad-mouthing people. It is rubbish in a sort of scientific sense. What we really need is, of course, these attempts are leading somewhere, but they are more like false steps, like how something is going on, something needs to be understood and rephrased. But what we really are facing is a new philosophy of nature where nature and technologies are becoming so intrinsically related that it's very hard to be a green person who thinks that the environment is only trees and valleys, a new philosophy of anthropology which I will discuss today and basically a new political philosophy which I discussed this morning with some colleagues at UCD. But there's plenty of work here that needs to be done because we are, as you were, stepping into a new territory. Now remember that philosophy never rewrites chapters, adds chapters to its book. So I'm not saying that what has been done so far has to be dismissed. I'm saying it's time to write a new chapter. Having read the previous ones, relying on the previous ones, that's a good story we have. But no, it's chapter 20 time to write chapter 21st century because that chapter is still missing. Time and high pastry, space. It's a beautiful reference to Galileo, local hero for people from Italy. Two, at least two lessons to be learned here. One, it's a very famous quotation which describes the book of nature in sort of mathematical symbols and no sciences reading the symbols. I said two, but actually three are the points here. One, a lot of science used to be considered by Galileo, Newton, until recently as a description of the world, but there's plenty of science there that makes the world, builds the world. You speak to any poetic, as in Greek, poetic is construction science, as in engineering, computer science. And they don't simply describe the world, they're actually building it. So I was not a scientist, well of course it is, you better not get a grip. But it's not a description of planets, it's a construction of, say, ICT system. So this divide already happening, which Galileo didn't have. Second lesson, a bit of philosophy of mathematics, for those of you, maybe for the Q&A. Funny enough, I mean, Descartes and the algebraic revolution has already happened a long time ago, but Galileo still thinks that the books of mathematics is a geometrical book. And for anyone who lives in the centuries, they're like, we don't think this way anymore. For us, if you ask anyone down the road and says, what's mathematics? Or two plus two. I know the numbers. They would not say geometry. But we were coming from about two millennia or so, or geometry being the queen of all mathematical sciences. The shift hasn't happened yet. It's a beautiful point. But above all, what does it mean to read a book when you actually have a science as I said a moment ago that can actually write the book? Oh, this is interesting because science is no longer a matter of reading what nature has written there on the book, but it's actually a way of adding to it. And this is what I like to discuss with you. Now, this Herbert Simon and Nobel laureates in a variety of topics as well, great for you also, but also great economists and so on. The beautiful book, if you have a chance, the science of the artificial. He writes, AI is not a science of nature. Remember, the book of nature, the description, read the book, let's say biology for you, or a science of culture. It's not anthropology, but it's a science of the artificial. And well, thank you, but so, what is that? Well, this is my contribution to clarify the point. I'm not sure what Simon would have agreed. It works well in English. It doesn't work in other languages. AI does not describe the world as, say, astronomy. It does not prescribe the world as, let's say, law or other similar endeavours, but it actually inscribes the world with new artifacts. So when you have the book of nature written in mathematical symbols, what happens if you write a piece of code? Well, you just add it to the book of nature. Well, that's the way of looking at this. So it's not a matter of building robots that can read the book of nature, because once you build a robot, you actually add it to the book of nature with a new artifact, which is a mathematical artifact, et cetera. Well, if this works, then there's a way of understanding artifacts in a way that has been sort of obscure by the debate for the past 50 years on AI. AI used to be a debate between two departments, cognitive science and engineering science. AI departments from cognitive science, they tell you, we want to create non-biological intelligence. It doesn't matter how stupid, but it will be the real thing. Maybe the intelligence of a spider, but it will be intelligence. Non-biologically generated. The engineering says, I don't care about intelligence. All I care is that it does, so and so, this and that, which if I were to do it, it would require my intelligence. But the thing there, it can be as stupid as a toaster. It doesn't matter as long as the problem is solved. Now, in this problem solving versus cognition, et cetera, what we've been doing meanwhile, these two blocks, they were still thinking about artificial intelligence as something that you would build and either being a problem solving or the real thing at some point, the real intentions. But meanwhile, what we did was to transform the environment to make sure that the environment would be stupid, DT friendly. It's our environment which has become AI compatible rather than vice versa. That's why Google's car can't work because there's plenty of data out there, sensors and databases and maps that can fit into it, not because there's a shred of any possible intelligence in it. So what I like to call the datafication or envelope in the world, that's what has happened really. Now, in robotics, an envelope, boring this from robotics, is the 3D space where an arm is successful, can operate. And think of it, what we do in industrial context is to build a whole environment around the stupid robots. You don't unleash robots in the street and say, build me a car. You build a so-called ontology in other context, an environment around the simple or perhaps sophisticated but still stupid abilities of the machine there. So the picture that I like you to have in mind to simplify is the dishwasher. The dishwasher is the perfect realisation of a robot. No, stupid mechanism inside, you build a whole environment around it to make sure that the stupid mechanism is successful, very successful. Nobody in his right mind would not wash dishes that way in the kitchen. This is what we've been thinking about for 50 years. No, someone like me doing the dishes like me and it doesn't work. No well, no nice. It's not there. And this is the future, an arm that puts the dishes in, sorry, because that's me at the moment. Someone has to put the dishes inside the ontology and the dishwasher and that will be a human interface. The human interface will come back in a little while. So while we were dishwashing the world, that's where AI started getting real traction. If you think that this is all science fiction again, this is the digital end of Europe, collects enormous amount of data on all possible things. They collected data about how many people in Europe use a laptop to access internet wirelessly away from home and work for other reasons. But to me that's the dishwasher. Where are they? They never left the dishwasher. You're not home, you're not at work, but you are constantly within the environment which has been created to make sure that your safe wireless connectivity works seamlessly between here, in Europe, et cetera. And there's a lot of people. I mean, more than 20% of all population has 500 million people here and growing. So in the past, grandma used to walk inside a computer. Now they're inside a dishwasher as it were. Then at some point her daughter stepped out and the computer became something that was in front of her eyes, but today her granddaughter is actually back inside. Is back inside in a way that she doesn't really see the computer around her, but all those sensors, all those things that are working, all that communication, I will show you in a moment they're all happening around there. So that's the space envelope in the world that we're living in. Big change in technology, big change in our age, big change in our environment. What about us? Which is always the question we ask anyway, so at the end of the day. What has happened to human age? And that's basically the topic of the book, The Fourth Revolution. It's the fourth because, oh by the way, still on record, John Searle, a philosopher, has criticised the book not understanding it, thinking that I was attributing the fourth revolution to myself. If he had read the book before writing the review, he would have known that I was not. There is a hero here coming up. Who is not from Oxford, I have to say. So the previous three revolutions were put together by Freud, self-serving advertisement, but no, smart guy, and he said, look, what I'm doing is something that has been ready done before me. I'm discovering a new area, a new scientific endeavour and so on, a new science that cast a completely different perspective on us. The first revolution, Copernicus, we thought we were the centre of the universe, all of a sudden we are moved outside that and of course it makes a big difference in terms of your anthropology. I mean, who you are, who you think you can be, how important you are in the universe. On this tiny little speck in the middle of nowhere, maybe that's not so relevant. So we retrenched and as a way we use a second special stage, which is this biological centrality. We were the species, and of course Darwin came and sorry, you have to give up also that centrality. And Freud said, well look, at least we retrenched into a sort of Cartesian centrality of the mind as being transpired to itself, rationality. Say, your brain is like a shoebox. If you look inside, you see exactly what's in it and normally nothing, but if there's something you will find it. And he says, well, sorry, you have to give up that too. You've got multiple you fighting with each other anyway. So these were three revolutions. And my suggestion in the book is that we're just undergoing a fourth revolution, which I attribute to Alan Turing. And he said, look, the real difference here is that by developing all this wonderful science, computer science and ICTs, we're casting a new light on ourselves. Not because this AI is really intelligent, but because it's telling us a different story about who we could be and what we could become. So we're not disconnected agents. No, pick up any book and there will be this single agent possibly rational, volume form, et cetera, with interest. No, it's all about networks. We are connected and we are kind of information organism. We've always been. I mean, we live by information in terms of expectations, fears, memories, feelings and so on. And we share as sort of information organisms, this infosphere, which I described before, this space of information with other agents. Some biological, some non-biological, some hybrid. And that's the kind of revolution that I explore in the book. So the difference is that in this infosphere as information organisms, talking to other agents of various nature, things like this used to be the norm. You had a barcode and luckily someone put a bit of English under it. But the new barcode of Wikipedia is not this. This is where you get. And what's the story behind that this is not meant for our eyes? As simple as that. This meant for artificial eyes, for other agents to be consumed. So we are generating in this world information that is not for other human beings, but it's for machines to read and consume and register and process. And you start thinking last time you had to prove that you were a human being to a machine because that machine was worried that you might be another computer trying to crack that particular sort of code or anything, including try just to generate another entry in Wikipedia. Well, that sort of movement or saying, well, show me that you are a real human being by reading this sort of mixed text and so on. But that's the world in which we live. It's not made of intelligent agents, but it's made of smart agents that can interact with us in different ways. So this is the picture that we have. There should be no sound, I hope, because no sound is meant. And it's just a little movie to show you how much we are sharing with the rest of the world. So this was 92 and there were about a million things connected with each other. Move forward. And 2003, a half a billion things talking to each other. It's not us, eh? These artifacts talking to each other. Then you start talking about the Internet of Things in 2009 because there's a lot of stuff about 8.7 billion entities that are interacting with each other independently of us. And of course the story goes on. It's going to be 11.2 in 2013. Further down the road. Keep going. Thank you. And you reached yesterday 14.4 billion. 18.2. For those of you too far away still 22.9. It ends in 2020, don't worry. 22.9, 28.4. And down the line, all these things that are interacting with each other. That's multiplied by billions the effects that you see when you go home and all those little red and blue and green lights going... But that is something talking to something else. So by 2020 we will have about 50.1 billion things talking to each other in this infosphere which we share with them. It's a bit confusing and unclear to know exactly what that means. So here is a fixed picture. This is not sort of projections by Cisco by the way. This now you can find it everywhere on the internet, so not secret. World population 2020, 7.6 billion people, more or less. Connected devices, 50 billion devices per person, about 6, 7 per person. Now that includes a lot of people who never made a telephone call in their life. So people in this room multiply this by 3, 4, 5. Make sure that basically you have about, for each of us, there are about 30 entities out there talking to each other behind our back. This is what it looks like. This is humanity and how many we are by 2020. And these are the devices talking to each other, how many they are or they will be. And in case you have a different perspective we are fast disappearing as the communicating sort of species on this planet. If you come from Mars, you want to talk about communication, well by 2020 this, make it 100%, these are devices and the line is going to go down, especially since we hopefully will stop multiplying ourselves whereas they keep growing. So scary, maybe, maybe not, but what are the challenges here? What's going to make a real difference? And I'm going to list five challenges and these are broad topics and I'm sure we can have the rest of the day to discuss. I'm doing fine with time, I'm looking at the boss. Yes, okay. First of all, this environment, is it friendly? I mean, are we really building this environment in a friendly manner? No, we had a wonderful conversation over lunch about how unfriendly or say less human friendly this environment can be. And I mentioned already in lunchtime that each of us has had the experience of being told that's the only way we can do it, sorry, because that's the computer way of doing it and yes, I understand, I shed a pain, I wouldn't do it myself, but I'm sorry, that is the way it has to be done. So how friendly is this environment? Sorry, how friendly is this environment to the world? Coming back to the environment, in this situation because of the gadgets we have, don't notice this too much, but growing up and I can see a few people like my age, there was a time when that laptop on your legs was really warm, unpleasantly warm. I mean, it was so warm, you had to have something in the middle because you couldn't just hold that laptop on your legs for more than an hour, it was burning. That's thermodynamics, that burning is coming from somewhere, that's energy. Energy is consumption, consumption is the environment. So this is recent data from the climate group 2008, apologies for recent, but they try to quantify the advantage and the disadvantage caused by ICT when it comes to environmental issues. And this is what you gain in terms of decrease, all that green stuff is not, as it were, is harm not done to the environment thanks to ICTs. Good old days, no true, but just as a joke, when we didn't print books but only digital stuff, well that's no, less consumption, good for the forest of that. But at the bottom, we forget that there's a black, dark sort of impact. Remember that laptop on your legs? Well, that comes from somewhere. And that's why, for example, all the energy, the green energy produced by Finland, the whole thing has been bought as a block by Google. All the energy produced by Finland, green-wise, it's going into supporting Google warehouses. And that's why we build warehouses next to a power station because it consumes energy. And therefore, what we're doing here for the chess players in the room is a gambit. We hope that by decreasing this much, while increasing this much, the impact on our environment, we have enough time for this to be a successful gambit. We are losing a poem, as we were, to win the game. If there isn't enough time, the poem will lead to more losses and the end of, as we were, life as we know it on this planet. So it's a gambit and I think it's a sufficiently reasonable gambit, but I'm not sure that the politicians who endorse this picture see it as such and understand it as such. I was telling you about a human-friendly point before and we moved through the environment. But the human-friendly idea is basically summarised. I actually borrowed this from a company. They had this beautiful presentation and I said, well, can I just use it? Sugetting. And they would shape our buildings day after day, shape us. That's so smart. I mean, it's really just true. I mean, there's no determinism in the kind of technology that we build. But once we have built them, they really stay there and they stay there for the next generations. The cab driver was complaining about the narrow streets in Dublin. Well, we didn't put them there for cars, but now they are there and then you better negotiate them. So how friendly is the environment that we're building? So against any determinism of any sort or things like the wide magazine editorial, what technology wants, nothing. Don't even try that trick with me. It's our responsibility. Don't for a moment think that, oh, it's technology. Yeah, as if. It's up to us to shape the next generation's world. And they will complain with us if we screw it up. Having said that, it is true that once you put it in place, it's very hard to reshape it from scratch. Why a Roomba world? I think that will be memorable. This is by the way someone who actually attached a little camera on top of Roomba. Roomba is, I see some Roomba. Roomba is, I'm not being sponsored, by the way, but I do have a Roomba and it's a little robot that cleans the house for you. It's a Hoover, highly recommended, especially as I was doing the Hoover. This big, this high, it's like a big, sort of thick Chicago pizza going around. There are several models, I have the cheapest one. They're very expensive. There's a Roomba also for the garden just in case that does the grass. But why a Roomba world? Because recently, moving house, my wife and I decided to have a new sofa. And we decided, of all the sofas we like, knowing what we're doing, we actually decided to have one with higher legs so that Roomba can go under it. And you start thinking, how many choices of that kind are we making today? I mean, this is a joke so that it sticks in your mind. But how many Roomba worlds are we building right away? And if you have a paying system where, as we had some point in a town, where the only way to pay for the car park is through a mobile phone attached to a credit card where just a week before it was coins, well, there's a Roomba world because you have cut off a whole chunk of the population. So the complaints are enormous and the city council has to scrap the whole project and go back to coins, loss of money, et cetera. Challenge number three. We only have five, so bear with me. Make stupidity work for intelligence, which is no easy because in the couple here, the analogy is between, you know, with the sort of spouse, spouses, and imagine that she, being on record, better be careful about wife, she is super smart, but also super lazy. And he is an idiot, but very accurate. He does everything, 24-7, et cetera. Now, the question is, who is going to adapt to whom? The intelligent lazy to the active, but total stupid, or the totally stupid, but very active to the intelligence. Of course, it's the intelligent lazy that's going to adapt. Oh, yeah, I didn't quite like it the way you do the dishes, but that's okay, you do them, I'm fine, and so on. So basically, you know, the idea of making stupidity work for intelligence is not trivial because intelligence has the wonderful capacity of being adaptive and stupidity, the wonderful ability to be tidal-less. And that's not trivial. So why is this an issue? Because recently, you might have heard, I hope not. But just in case, you were astonished hearing some famous people from Cambridge at the other place. What do they know there? About the arrival of AI, I mean, and they will dominate our lives as a total nonsense. We've got plenty of problems, but not this one. Now, this is Alan Turing, that's the Lebanon Prize, and that's the medal that they give to the Lebanon Prize. The Lebanon Prize is awarded to any piece of software that will pass the Turing Test. The Turing Test being something just for those of you who haven't been exposed to this very quickly, you are in this room, you interrogate two entities on the other side of the wall, you don't know who is who, and by a question and answer game, you need to guess who is who, who is the machine and who is the human being. If you cannot spot it, there are a few constraints in such and such amount of time and so many questions. If you cannot spot the difference, well, the machine has passed the test. It's as good as the human. Hopefully, you pick it up an intelligent human, because of course, if it is your moronic neighbour, well, you won't see the difference, trust me. But hopefully, you have a brilliant guy and the machine, and if you don't see the difference, pass the test. Now, the Turing Test therefore has become a kind of a distraction and a bit of a game and so. When Turing introduces the Turing Test in his famous paper in Publishing Mind, 1950, he asks a question, which is, can a machine think? And then he adds immediately after, too meaningless to deserve discussion. That's not a point. You don't ask whether a machine can think. You ask, can he perform as well as something else that we know can think? At that point, we know that is passing the test. Unfortunately, on the Lebanon Prize, there's exactly the phrase that they printed on the medal, can a machine think? So either there's a very subtle irony here and we didn't catch it or they didn't read the paper. I went for the second option. Who reads these days? Now, this is the same problem multiplied by a few billion dollars in California, Google. You must have read also that Google is buying anyone and anything that does artificial intelligence on a global scale. Seven companies last time I counted. There's audio, but we don't have time and I don't want to show it to you, but this is the text for those of you too far away. It was an interview that Eric Schmidt, at the time executive chairman of Google, gave. It was a meeting at the Aspen Institute July 16, 2013. That is important, 2013. Ask, will any machine will ever pass the urine test? The urine was deadly wrong. It said, oh, in 50 years of bone or 30 years or whatever, it gave a number and it was like no even close. But Eric Schmidt said, well, quote, many people in AI believe, that's why you are the executive chairman. That's smart. You don't say I believe because then I can prove you wrong. He says many people believe, oh, it's a problem, but that's. Many people in AI believe that we are close to a computer passing a urine test within the next five years. And I hate aubergine and on record, I bet a plate full of aubergine that is not going to happen in 2018. By 2018, hello. I will eat a full plate of aubergine if this is going to happen. Why? That's an easy bet, so it is the actual interview. The easy bet is based on this. This was the winner of 2013. Oh, by the way, three prizes, gold, silver and bronze. Two have never been awarded to anyone, the gold and silver, which is their both science fiction. The bronze is awarded every year to the less-worst performer of all the bunch of applications. The one that is less rubbish wins this few dollars. It was Mitsuku Chatbot. Oh, some of these have weird behaviors when it comes to sexual interaction, so you have to be careful. He won the bronze medal for the following conversation. And again, I really, for those of you too, far away, you can try it online. Me, what can someone do with a pair of shoes? Replyd quite a lot of things. OK, that's fine. I don't know the difference, such as, for example, a tomato. Well, that starts being weird. Anything else? That is all I have for right now. Remember, I need to spot who is the computer and who is the machine. What's wrong with the following sentence? The capitals of France are three. Lyon and Marseille. How can there be self-help groups? So the usual trick of these machines, you know, replying to a question with another question, I tried that with an undergad and the examination never works. But are you sure? Well, I used to be indecisive, but now I'm not so sure. Thank you, computer. You are quite welcome. What sort of computer any go on and on? And this is the sort of interactions you can have. When you say, oh, that's good news. That's all sold. When the last London Prize 2014, that piece of chunk got the third in line. This is a conversation. And when asked, this car couldn't fit into the parking space because it was too small. What was too small? I don't know, I'm walking in Cyclopedia, you know. Because this is as far as we went 50 years ago in this sort of fake interaction where you can just talk in automatic pilot and I've done this a trillion times at a high table when you don't want to listen and say, what's your topic? What are you doing? Oh, it's a watching college. You just go through the usual stuff and you know even listening. Now, of course, machines can do that for us, but that's why all this came in 2008, whereby we thought that we were getting stupid by the day and machines were getting intelligent by the day is past its sort of headlines interest. The end of theory, the data deluge makes the scientific method obsolete. No, it makes this article obsolete. And it has been obsolete since Francis Bacon decided that all you had to do in science was to collect data. Wrong at that time, wrong for the past 500 years, nobody does science in this way, nobody. Collecting data and asking questions. It's the first thing you say to students, don't go to the library, don't read, think, come up with a problem and then you check the data and then, or is Google making us stupid? What the internet is doing to our brains? Now, remember now, machines getting more intelligent Well, no, what Google is doing, or things like that, it's polarising. It's polarising because the stupid become more stupid and the intelligent become more intelligent. That's the trouble. Because if you're smart and intelligent these days, with those tools you can do amazing things but you have to be pretty smart. But if you're dumb, well then you become dumber by the day because of course you relax. The analogy I had in New York during the festival I said, look, it's like a car. You can take the car to go to the gym and get fit or you can take the car to buy the milk and get fat. So, polarising. Non-neutral, is that the mistake? Oh, so, oh, it's not neutral. It means that more will get more and less will get less and that is the real trouble. So, problem solved? Not quite because of course these machines being smart but also stupid. They need humans as a module sometimes. So here, these are New York Times articles and some time ago when doing American elections, Twitter started getting problems in disambigrading the phrase, big bird. Because of course nobody was referring to Sesame Street. Everybody was talking about the US government, big government and so on. But they used big bird as an analogy. And Twitter came out with something that everybody knew but stated loudly that they use graduate students to disambigrate this stuff because they're the only ones who can actually grasp the meaning and make sense of the meaning. At this stage, or maybe in the future, there would be other machines doing it differently for us. But what's happening here is that you have memory outperforming intelligence, meaning algorithms and big data. Now, basically doing better than you remember last time you played chess and you won against your iPhone. Good old days. But synthetic engines, namely our computers, they need semantic engines like us. And therefore, we start becoming, as you were, human inside module within a bigger machine that needs us and our intelligence. So what has happened in the past is being, a phrase I like to use is that now, future is where the past happens again with a twist. And we often work as interfaces. Remember the arm that I show you, between the dishes outside and the dishes inside? Well, I'm just the interface between a dirty kitchen and our cleaning up system. And now, hopefully, I would like to have a robot there. But now, if you look at the gentleman or the lady here, they are just interfaces between a petrol station and a car. And we could not think at the time, and we did not have the technology to make sure that refuelling a car could be done by a robot. But for goodness sake, I mean, we have robots on Mars, but we cannot refuel a car automatically. You need a human being to do it. To unhook, put the cloth and put it back. If you think of it for more than one billion cars in the world, for I don't know how many petrol stations, if anyone invents a system doing that automatically, that's the next billionaire. But it's very hard. This is a strange robot in all different models. Are we going through the same problem with electric cars? Yes, we are. Have you seen that? There's someone who has to plug in the thing. No, don't do that. I mean, clearly, if we're going to go electric, surely there has to be a way of parking and then you leave it. I don't want to be the one who actually plugs in the thing and then one billion cars electric and ten years later just have the same thing. Oh, if only we had thought. But the lady is the same problem. She is an interface between a GPS and a car and she's also fast disappearing. The gentleman on top never met anyone like that, but my grandfather would have. Or the lady at the bottom. This is something that we often forget in terms of interfaces. Technologies help us to remove interfaces not in the sense that we're not going to do it, but in the sense that there's no job there for someone to do it for you. I don't know about you guys here, but since we're back in Oxford, it's me scanning all this stuff all the time. So it's me putting the stuff in the dishwasher, it's me scanning the chicken at the counter because the job is gone because the technology is gone good enough to make sure that you don't need a specialised human to do it. Anyone can do it, therefore, the customer can do it, therefore, it's up to you now. And I'm not being paid to scan my own food. Which leads me to, we're still on that particular challenge. Don't stay with me. This is recent data about the American job market. I think you know where I'm going. How are we doing with time? We're going to continue, aren't we? Oh, yeah, we're doing quite that. Again, for those of you a bit too far away, I apologise for the graphics a little bit too small. This, of course, as you can tell from the economists, they divide, I don't know why, exactly between services and government, probably because the government provides a disservice, I don't know, I mean, whatever it is, the reason. But between services and government, if you consider the whole population of the job market in the States, that's 90 plus percent of jobs. Services and government. Manufacturing is going down quickly and is already way below 10%. Well, remember, more than 90 below 10. So just a fraction, two, three, maybe four percent, depending, is agriculture. And someone, when I said this recently, he says, oh, yeah, but not in terms of GDP. No, no, no, also in terms of GDP. Agriculture doesn't contribute much to American wealth. And if we could just stop having that hang out from millennia of starvation in Europe as well, we could do much better. But what's the point here? Well, very few people have anything to do with bioware, agriculture, just a little bit have something to do with hardware manufacturing and most have got to do with software one way or another. This is fine, it's a bit weird, but well, that's what happens, and you must have seen this before, what happens when you start calculating the effect of computerisation on the job market. If everybody is in the software industry, more or less, and handling paper and handling information one way or another from airlines to, well, if you start an automatisation of their market, that's the kind of curve you're going to get. Some jobs are super safe. You don't want to get a massage from a robot. It's just not going to happen. So that's going to be safe. But do you care who actually handles your flight tickets as I say at the desk? Oh, you couldn't care less. So those jobs are going to go. And basically there are 47, about 50% more or less of jobs at risk according to this particular calculation which was done as a study in Oxford to make news and so on. I said, okay, well, that's an advance. Society maybe, maybe so maybe not. What is missing from the picture? That is more data. Well, recently I was in Finland for an interaction with the Ministry of Transport. And interestingly, in Finland they run the same analysis. But with one variable that the previous analysis hadn't taken into consideration in full, which is legislation. What if the legislator says, yeah, that's doable, but we don't allow it? And if you think that that's just a joke, imagine the following. How many trains now in Ireland are going from one place to another as a fixed road can take turns without a driver? And the answer is zero. To the best of my understanding, there are at least in Britain, a train is not allowed to go from London to Manchester without a driver on it. Why? Because of the legislation. I mean, you just don't do it. How many airplanes are flying from one place to another without a pilot? Well, this is all doable. I mean, it's all, especially trains. I mean, nothing more than going from A to B on a train. Well, the legislation doesn't allow it. So how many cars are going to drive downtown without a driver? Say, how many cabs are going to be there? Well, it depends on the legislation because if the legislature says illegal, that's the end of the problem. They will not happen. So it's not about technological feasibility, it's about the legal framework that allows that or not. Once you take that into account, this is the American picture, same data, just a different graph, but exactly the same. And that is on the left hand side, the Finnish picture, which I will say a more European picture where basically Brussels and Helsinki have been taken into account. What happens when the legislators say, we don't like this, you better have someone on board. In fact, two on board. In fact, how many stewards will those three people exert? So the future is complicated. And when you're told that jobs will go, they will go if technology is left alone, but we are also in charge. And depending on the legislation, they will or will not go. It depends. It's an open call. You may say it one day that supermarkers are forced to have people at the counter to scan the stuff and build it for it. And if you do that, well, plenty of jobs around the corner. China, number four, and we're getting close to the end. May predictability empower freedom. Remember those were the two challenges we've been exposed to. Smart technologies are becoming really good at doing things that we do because of our intelligence, but also getting good at predicting our freedom. And that's one of the two pillars of our definition. We are free, intelligent agents, different from humans. This is, you don't have to read it. It's a passage from DECA that has kept hundreds, if not thousands of students in Oxford awake at night trying to understand it because he says something really odd. He says, suppose you measure freedom on the most free of all possible agents, God. Nobody can be more free than God. Well, God doesn't change his mind. It's not that God thinks I'm going to shave today as opposed to I'm not going to shave tomorrow. He's not going to make a different, no. So God is rational and is free. So the more you incline towards a particular rational free decision, the more free and rational you are. You just ask free and ask so rational as God. Couldn't be any better. That was the model. But unfortunately, that is highly predictable. We all know that. The closer it gets to two plus two, the better you get at saying four because that's the only. So if you are very rational, incredibly rational, as we were saying on lunchtime, you're going to come back home with the same toothpaste, even if you don't remember which toothpaste you like. And that's an experience that's rather disturbing. So if you're very rational and basically that's what you go through, this is a famous case. You must have seen it before. So it's just a reminder of a target developed in 2012. This is an old story. Developed a whole series of product, about 25, which show that if someone buys those products in their order at their time, not only is she pregnant, they know at which stage of pregnancy the person is, months. They just forgot that maybe there was a bit too personal. They sent tokens for their pregnancy and the tokens were received by the father of the lady. It was hugely upset and complained, called target. It says, are you insane? Are you crazy? My wife and I, we don't know you're having a kid. I said, no, we're not talking about your wife and yourself. We're talking about your daughter. So my daughter is only 16 years old. I had an explanation and they always said that she was going to tell them sooner or later. The target already knew. The family didn't. And luckily that was in the United States. She didn't get stoned or thrown out of the house. I said, okay, I'll find one more member of the family. But this is the kind of technology that is available today in terms of challenge. How do we make this technology improve, support, enhance, foster freedom as opposed to become a constraint and predictability becomes something that we carry around our neck as a way. This is real stuff. And of course it always happens in the States. Since 2008, old news. Of course have been the police and of course adopting predictive policing and predicting measures. This is just the data again, economy stuff, no research, no nothing else, only he knows. No, it's public domain. This is what happens in LA when you adopt some predictive policing. In other words, you start saying, well, some crimes have happened there. We think they're going to happen again or not. What's the probability and so on? The place where they adopted it and the other places where they didn't. Incredibly successful. It does work. I mean, we are predictable, especially in our large numbers. The larger the numbers, the more predictable we are. And fun enough, that's what I was told, not that we know. The software that has been used to have this predictability of recurrent crimes is the same software that's California that was used initially developed for earthquakes. The point being that they wanted to know where earthquakes happen again if they have already happened once. And no, the maths is the same. If something happens, how likely is it that it's going to happen again? And you'll change the parameters. It's crimes, not earthquakes, but bingo. I mean, these human beings is not Earth moving, but you have the same solution. So the effectiveness tested by the LA police is huge. And to the best of my understanding, this has already been adopted in the UK in Manchester. So it's getting some traction. But above all, that was the surprising thing. And again, let me just read for you. This is February 2013. So it's still recent, but not that recent. It's not yesterday. It was the Philadelphia Court beginning using computer forecasts to predict future criminal behaviour and change therefore the penalty. So, well, this guy is going to do it again unless we really sort of make it sort of stick as well. So just a little bit more punitive in case. Or I said, no, no, this is just a good guy. He's not going to do it again, so we can be more lenient. Based on what? It's not minority report. Based on ordinary forecasting, data, stuff. And yes, it's already happened. So this is not, as we're technology, helping us to be free, but us sort of linking freedom to. And because we are very predictable, this is where some of the challenges are coming from. Remember freedom and predictability. This is a quick story of Nike with some highlights. 2006, Apple and Nike launched Nike Plus iPod. Make a couple of mistakes, excuse me, mistakes in terms of privacy. They fix it and then start again. So running tracker, Nike starts shifting the branding strategy. Nike used to be a producer of things, but they want to be producers of products that are experienced. So processes that are not stuff. There's no real money in producing stuff. That's what China does. But experience, that's the real money. So in 2012, Nike starts producing the fuel band. There's more stuff. But you start thinking, oh, what's the real money? This stuff, 2013, 18 million users of Nike Plus online. That's the real money. We've got 18 million people who have money to waste to buy a fuel band. And time to waste, credit cards, probably English-speaking, has some mind-gold. 2014, 28 million users of Nike. Nike starts saying that they will discontinue the fuel band, concentrating on consumers' experience, elementary. But what is the consumer experience here behind? Is the predictability of these consumers and what you can sell them next time. I don't want to picture this into dark tones, but unless we are aware of what's going on, I think measures will not be taken. And finally, looking at how we do it. Can, it's just a quarter past now. I know some people have to get back. If you could just summarize, because I think it takes some time for questions. We need to make technology make us more human, no less. And instead of showing you what tattoos look like these days, or the kind of funny things you can do to your arm, this is real stuff, not by hacking or growing on your arm, implanting things in your brain, or guiding things around. Basically, we need to make sure that the technology that we are developing, as I said, we are in charge, is going to make us more human, no less. Remember the Roomba world? Well, that's part of the pitch. So what can be done in conclusion? Well, I think we need to do this a couple of things, the two, I as you were, one plus one, but it really is intelligence plants inventiveness. We need to be smart, and no, smart doesn't technology smart, but we need to be intelligent, and we need to be inventive of our solutions. That's the only way forward. So bottom line for the financial among us, we need to upgrade our views about who we are, what the world is like, and how we can track among ourselves and with the world. That's four things. Us, the world, interactions among us, interactions with the world. These are the four lines where we really need to upgrade our philosophy. And it's a new chapter, it's not erasing the old ones. It's not burning like Hume said, the books in the library. It's just writing a better chapter for us. And for that, I believe we need better, not less technologies. I'm not anti technology in any possible way. We need to have smart technologies, absolutely. We need to have technologies that regulate other technologies much better than they do now. You want to have some technology checking that other technology is doing its job properly. And we need technologies that monitor other technologies so that everything in terms of what's happening and how something gets regulated is also monitored so that we can step back and have a better life. But for that, there's a lot to be done. And thank you so much for your attention.