Watch videos with subtitles in your language, upload your videos, create your own subtitles! Click here to learn more on "how to Dotsub"

David Wood - Technological Unemployment - London Z-Day, 2015

0 (0 Likes / 0 Dislikes)
All right, I hope you all had your lunch and everything, and are feeling raring to go! So, I will introduce our next speaker. Okay. So, David Wood is Chair of the London Futurists. He spent 25 years envisioning, ... architecting, implementing, supporting and avidly using smart mobile devices, including 10 years with PDA manufacturers Psion PLC, and then more with ... smartphone operating system specialist Symbian. His background includes many years building and integrating UI system software and application frameworks, and directing technical consulting teams, with the lead ... phone mobile manufacturers in the world to create the first successful smartphones that you all have in your pocket. (or hopefully) In 2009 he was included in the T3's list of 100 most influential people in technology, and he's also executive director of Transpolitica, an activist think tank founded in January 2015, to accelerate the adoption of the better politics of tomorrow. So, please welcome to the stage, David Wood. [Applause] So didn’t we have three very good talks this morning? I was personally familiar with all these subject areas, having been at a few Zeitgeist events before. But I thought these three talks this morning put things particularly clearly, particularly vividly, so I’ve got lots have new reading to follow up on from that. And these three speakers all looked at key problems in our present day: issues that are demanding answers. I’m going to look at a different kind of problem. It’s related as we'll see, but it's a problem which is going to be brought about by the rapid progress in technology and how it's going to change the nature of work, the nature of employment. And don't laugh, I’ve given it the title ‘Anticipating the rise of robots, ’ well there’s a more technical term for it: technological unemployment. And today this is a bit of an issue, and as I am going to explain this is going to become or more of an issue that's going to require the same kind of zeitgeist change to address it that we heard about in the morning. And although I've called it a problem, it is also an opportunity, as I'll describe. So to start off with, let's remind ourselves the summing of the relentless pace of change over just the last 10 years in terms of what’s happened in the world of computing alone. You may have heard of something called Moore’s Law, which summarized: We can say that over the course of 10 years, computers get 125 times more powerful, over the course of 10 years, for the same price, for the same size. Or you can do it differently. You can use this ... improvement in computing technology to get a single computer which runs 5 times faster than before but was also 5 times cheaper than before. So if it cost you 100 pounds 10 years ago, it will be down to 20 pounds today, and go 5 times faster, and also be 5 times smaller, which means it will take a lot less battery power. And also, computing can store more information. And then there are improvements to networks. You've heard of 3G and 4G and 5G; there are better WiFi networks the whole time so data can be shared more. So it's not surprising that many things have changed in our personal computing in the last 10 years, many things we now take for granted, but which 10 years ago almost none of us were thinking about, or using. Things like electronic books. Amazon announced just a few years ago that they were now selling more electronic books than physical books, onto Kindles and other devices. And then there are tablets - tablets which have been taking over the world, a hundred million of them were sold in just a few years, changing the lives of grandparents, grandchildren, many professions, and many other people. We've seen the growth of smartphones, so that many people around the world, even people in comparatively poor environments in third world countries, are having their livelihoods changed and enhanced by these devices. And then there are applications. There’s something called Facebook. 10 years ago you needed to be a student at a few universities in North America to have a Facebook account. Now, probably most of us love and hate our Facebook account; many of us who spend a great deal of time with it can hardly remember what it was like before. And then there’s Twitter, which is also less than 10 years old. And by the way, we do have a hashtag for this conference today: it’s hashtag zdayUK15. So if you're getting bored with what I'm saying, then by all means have a quick look at what others are saying about this conference on #zdayUK15. But I'll try and keep your attention here. Even more remarkable I think, is the fact that 10 years ago, there was almost nothing on YouTube. I think there was about one video that had been uploaded, just 10 years ago. But now we take it for granted that we can send wonderful pictures of cats to each other all the time. And in case that's flippant, let's also remember what else YouTube has given us. Things like Khan Academy: a tremendous free resource of wonderful educational material, accessible by all, so that even the richest person in the world, Bill Gates, has his own children doing lessons on Khan Academy, not because they're free, but because they're just so good. And of course Bill Gates puts a lot of his own personal money into the Khan Academy. And then there’s things like Wikipedia, which 10 years ago had a few hundred articles in it, but, since that time, has become a tremendous resource for figuring out, in many cases, what’s true and what’s reliable. Zooming out a bit more, if we go back 20 years ago, here’s the kind of stuff that a guy called Barry Ritkoltz, a United States blogger, was using in his daily life. And it’s a bit of an exaggeration, but not much of an exaggeration. He says this is what he used 20 years later instead of all that stuff: just one device. So he no longer needs to carry that heavy camcorder around, that separate Polaroid camera, a Sony Walkman, a Palm Pilot, and so forth. And as I said it’s an exaggeration, but it shows what remarkable changes have been happening. And there are remarkable changes in other fields too. Just very quickly, some of you may recognize this city. This is from just over 20 years ago. If you look closely you might see some Chinese writing there; this is the Shanghai famous Bund waterway, and 20 years later it looked like this. So it is remarkable what we humans can do if we put our minds to it. Just imagine if we put our minds to more constructive tasks. And of course the Chinese can do this even more quickly now; they can build one of these skyscrapers in a fraction of the time. Now, there’s one view that says “Yes, we’ve had remarkable change in the past, but it’s sort of coming to an end. All the big things that need to have happened, have happened.” And here I’m quoting a view described by Kevin Kelly, who is a very wise writer on technology, and cofounder of Wired, which some of you may read from time to time. And Kevin Kelly refutes this view. He says “relatively speaking, nothing big has happened yet.” We're not even at the beginning of the changes that technology is going to bring us. “We’re just at the beginning of the beginning of all these kinds of changes." And he says that “The next 20 years are going to make the last 20 years just pale” in comparison. So if you think there's been a lot of change in the past, you can expect a lot more. And that’s a view I certainly share too. And I'll discuss this abstractly for a bit, and then I’ll move in and try and make it more vivid, and more realistic. In the big abstract sense, what we're going to see in the next 10 to 20 years is something called NBIC Convergence. This is not a Canadian radio station - it stands for NanoTech, Information Computer Technology [ICT], BioTech, and CognoTech. The stuff I've been talking about already is all in this corner down here - the bits mainly, when we can store things much more cheaply, we can compute things much more cheaply, we can transmit things much more cheaply. But there’s [3] other corners to this remarkable transformation. NanoTech is about manipulating things at the atomic level. And we can see a bit of this in 3D printing, when we're able to create things in more wonderful ways than before. But increasingly we’re able to create bots, nanobots, which are computers at the molecular level. BioTech means we're able to improve our genes. Scary stuff: lots of potential for things going wrong, lots of issues there. But if you do it right, we can become much healthier, much stronger, than before. I just saw today that Superdrug (may not be your favorite store, they do lots of good things), they are now offering 23andMe® genetic personal kit for you to buy and apply in your own home, to find out more about your own genes, which may be a good thing, may be a bad thing, but the change is coming. The last quadrant here is the quadrant of CognoTech. We are understanding the brain cells, and the neurons, and what's going on in our minds, much more fully than before. And roughly speaking what's on the left side of this big diagram is stuff in the physical world, the right side is biology. Roughly speaking, the top half is hardware and the bottom half is software. But the real change will come from the connections between all these different areas. Realize - we’re getting better and better at computing; we can actually design physical things more clearly such as 3D printing as I said. Hence, you may have heard the strange phrase that nowadays hardware is becoming software, that many of the remarkable changes which have taken place in the world of digitized information, will take place in many of the other manufacturing goods too. We can also have sensors, which are smaller and smaller. We can put them on the body or inside the brain, we can understand what's happening in the brain more fully, and once we understand how the brain does its remarkable tricks, we can then improve our own software, and so on it goes. So that’s fairly abstract - let's make it a bit more concrete, in terms of four big stories I think we're going to be reading in the magazines and online over the next 10 years or so. They're already here: 4 great convergences. The first convergence is between human and machine. We've had computing devices on our desks and on our laps, we've had them now in our pockets, we are increasingly wearing them on our wrists, in our clothing. We will have them, I believe strongly, we will have them in glasses [in] some stage, in the next 5 to 10 years, which will be extremely powerful, and then they’ll go inside our body; they’ll get smaller. And we’re already putting some things inside our body. My mother has put a new hip, courtesy of the NHS inside her body, transformed her life, and more recently she's had a new lens, solving some cataract issues in her eyes. In both cases she was fearful, but now she's very grateful. That's just the beginning of the changes we’re going to have. So increasingly, we are going to become not the six billion dollar man - because that's far too expensive - but as this technology becomes cheaper and more accessible, we are cyborgs with computing inside our bodies and outside. The second great convergence is when we apply more of our software skills into the world of biology. And instead of just coding in the world of silicon (semiconductors), we will begin coding in the natural language - the natural computing language - of DNA. So we will be creating in some cases, improved life forms, which can do remarkable things, hopefully, like taking CO2 out of the atmosphere. Can be improving the waste and reworking the waste that we generate. And we can also have genetic engineering. Not just for the scary story of designer babies, but also for the more interesting story in many ways of redesigning ourselves, that if we are aware, from 23andMe or other things, that we have some particular limitations in our biology, they can be fixed. It's a big story; let's not go down there today - let's move on. Third big convergence is between the physical world and the virtual world. You're all here today watching it in the physical world. But in the future more and more people will be watching these events virtually too in real time. People will be watching from their home, from all around the world. And they will be living in helmets which makes them think that they are here, and we will imagine that they are here too. This is a transition from virtual reality to augmented reality. We will have these helmets, we will have glasses, which will give us more and more useful information as we we’re walking around, helping us to enjoy our lives better, to understand what we're doing. And there are other things we could talk about, like the massive online open courses which are often free, which gives us the information like Khan Academy and many other courses, which is transforming education. Everybody complains that education is very expensive these days, 6,000 dollars, 9,000 pounds (sorry). In the future it will be much, much less, because we'll get much of the learning from these online courses. The convergence I most want to talk about, the one that's going to bring it back to the main topic of today, is the convergence between artificial intelligence, what computers can do, and the kind of intelligence that goes on in our heads: human intelligence. And my story is, and what we’re going to see increasingly over the next 10 years, is more and more capacity of these machines, robots, software, call them what you like, to do deep machine learning, to figure out by themselves what's going on in their environment, and to guide us, and to take decisions for us, or to suggest decisions for us. I'll give examples. And the way forward here, is that we can adopt some kind of hybrid intelligence, in which we use our own native wit, but supplemented by the intelligence that this new technology will provide. The even more scary thing is what happens in that middle there, when all four convergences are themselves converging. But, let's leave that for another day. I just want to pull back a bit, and say there is strong evidence in my view that technology is accelerating through positive feedback, more and more people are learning about technology, and universities, at homes, all throughout the world. More and more people are taking up some of these things in hacking formats, or as part time, or as a second or third career, and more and more people are building on top of each other's insights. And so I think it's fairly likely that in the course of these next 10 to 20 years we will have increasingly enhanced humans with extra intelligence, extra health, extra longevity, extra experiences, and to sum it up, with extra opportunities. And I imagine many of you have got lots of question marks in your mind too. You're probably thinking “Well hang on, there's another consequence to all this technology," which we can sum up: that it's disturbing humanity, because there are many things that can go wrong with all this extra firepower that we are putting at our own disposal. And you can look at it from different angles. You can say: well there are already some really angry and alienated and frustrated people in this world - young people, middle aged people in various terrorist groups or various militaries. If they got a hand on some of this technology, it could be even worse than what's already happening. And there are governments, some more benign governments, some very terrible governments, who can use this stuff [for] surveillance in one sort or another. And there’s what we're doing to the environment, and the issues on the climate, which I think, really the big existential question of our time. Then, there’s the risk that as we put more intelligence into robots of various kinds, they're going to unintentionally do things that we programmed in, and which we didn't foresee. Whether it's battlefield robots running amok, with various drones making their own decisions which come back to haunt us. Whether it's financial robots meant to be supervising the stock exchange and crash it and burn it, then we have a much worse problem than happened in the past - well we'll come back to that. But to summarize that, there is existential risks from this growth in technology too. In other words, this positive, accelerating technology cycle will benefit individuals - it’s got a strong possibility to do that - at the same time, it is threatening society. My big question is why aren't more people talking about this? If I look at what the politicians are talking about in the run-up to the general election, I see almost nobody raising these issues. It's a real issue of people being trapped in their own inertia, their own momentum, their own vested interests. Well, very glad to see that people in this conference are tackling some of the more fundamental issues. Some reaction to this is to say “Oh it’s not gonna happen. Technology is gonna run out of steam. We have had all the benefits that are going to come.” I think that’s wrong. You look around - you can see so much evidence of incredible technology innovation. So I don't think it's possible that technology is going to stop all by itself. Other people say “Well hang on, we'll just ... switch the switch off." You know, we're going stop people developing technology - that's how we’re going to avoid these problems. I don't think that's going to happen either, because there’s so many vested interests who are benefiting. And we actually are benefiting in many ways from these enhancements, whether it’s better hips, better ... lenses in our eyes and all this other stuff. I think there are however, two credible scenarios that people can take, regarding these stark issues facing us. And one of these viewpoints is quite common in Silicon Valley. It's quite common in a lot of big industrialists. It’s quite common in a lot of startups, and they say “Let's just focus on allowing individuals to enhance themselves. Let's not stop these technological companies, let’s encourage more smart research and development. Let's have free enterprise do whatever it needs to do. Let's keep the government out the way because the government [is] clumsy, the government don't understand new technology, the government regulators are always solving yesterday's problems instead of tomorrow's problems. Get them out of the way and we will improve humans.” And, they say ... (somewhat naively in my view), they say the world’s social problems will be solved as a byproduct. For example, this new technology will be able to suck the CO2 out at the environment, this new technology will have better solar conductors than ever before, better energy storage than ever before, and so that's how we’ll solve some of the issues of climate change. This view’s got a name, called Techno-libertarians. They love technology and they say “just let us free to progress as we wish.” As I’ve said, I think this is a naive view, it’s a politically naive view. It doesn't recognize that there are many issues that will not be solved just by technology alone. That we do need, as well as this research and development, we also need ... wise regulation, not to stop this, but to steer it, and smart governance. And although government hasn't made a great job in the past to doing it, we who do understand these technological issues - the benefits and the risks - must get more involved. And more than that, to change the social systems. So this viewpoint is called Techno-progressives, and I’m very happy to label myself as a techno-progressive, even though I think there's a lot of smart and clever and well-meaning people on the other side of that dimension. Now you may be thinking of all these existential risks, one of them doesn't really belong there. You may say "A robot uprising? That’s a bit of a, kind of a Hollywood scenario, isn’t it?" So that's what I want to talk to you about next. I want to convince you that this actually is something that we should be considering, on roughly the same level as all these other issues down there. And I'm going to show you four people who should be fearing the robot uprising. Number 1: there's the elevator operator, when he used to be employed to go in and take people up the lift and down the lift. When I was young I remember seeing these guys in various department stores. They've all gone now, because what? Lifts can operate themselves. It’s not a very clever robot - the user comes in and presses a few buttons, and you don't need a human operator to solve it. Number 2: the bank teller. You still occasionally can get money from a human being at the bank, but much more often you go and interact with a robot; it’s called an ATM - a hole in the wall. Number 3: This guy’s looking a bit concerned in there, I think he should be. He is the till operator; he's increasingly going to be put out of work by another kind robot - the check out. And they’re not very good yet, you can sort of get by these checkout tills, if you're a bit patient, and often you need a human to come and sort you out. But guess what? They're going to get better. They're going to get 125 times better over 10 years, which is probably going to be a lot better than most of the checkout assistants. Number 4 I've got down there, is the passport checker - this angry passport checker is in many cases going to be replaced by iris checkers and fingerprint checkers, and occasionally they'll be an interrogation that you need. So all of these people either have been put out of work or will be put out of work by robot uprising. A whole class of people. This is what used to happen in many companies. When I did my internship in a company in Edinburgh in - goodness - 1977, there's a whole bunch of people who looked a bit like this - the typing pool - and they've all be put out of work by Microsoft Office and the other software. So they've all gone too. So we're talking now about this technological unemployment, and I can summarize what's going on here by this Dilbert cartoon, in which Dilbert’s boss wants to get the admin organizer, Carol, wants her to organize a meeting. So Dilbert’s boss strides up to Carol and says “Schedule a meeting with Dilbert and Alice for next Tuesday at 10.” And before Carol can waken up, Dilbert’s phone says “Done!” “Never mind, my phone has taken care of it. ” And by this time Carol has woken up and is getting a bit worried. What’s she going to be doing with the rest of her life? I’ve been staffed out of a job. You may have ideas, you may tell her to go and retrain, become a software engineer or a ... taxi driver, or whatever. Well we'll get to that in a minute. Then there’s all the people who are driving cars. And this is a self-driving car which for many years people said was an impossibility. There were lots of papers written about how computers would never be fast enough to make sense of who's crossing the street. Is that a pedestrian coming into the street that you need to slow down for? or whatever. And initially cars were terrible and they would soon stop, they would get stuck in bends, but for the last 10 years they've been improving remarkably. And they’re going to be much safer than human drivers, which is why increasingly, we’re going to say: let's have computers driving these instead of the ... distracted drivers. But what’s going to happen then to all the taxi drivers? What's going to happen to all the lorry drivers who spend all their time driving freight up and down? Now you may think it's a bad use of the environment to have all that freight being driven anyway. But let's say, there are many people whose livelihood comes from driving today. And if you think about it further there’s probably lot’s of other people going to be made unemployed because of this, like the insurance sales people. "I’ll insure you in case you have a crash when you're driving!" "No thank you, I won't have a crash when I'm driving; I’m not driving anymore." And then there are managers. And this is an article from the Harvard Business Review from a couple of days ago, from the Institute for the Future. And they say ... "Fortune 500 executives spend a fair amount of time thinking about how automation and the Internet are changing the nature of employment." Maybe they don't need admin staff anymore. Well, guess what? They rarely wonder about how technology will have an impact much closer to the home, on their own jobs. And the people at the Institute for the Future, they wrote this article ‘Here’s How Managers Can Be Replaced by Software’ and there’s a demo app called iCEO. I think it does more of a project management role than a CEO's role to be frank. But if you look inside what it does, you wouldn't want to change this straight away, you’d want to test it out, but they have been testing it out in their own use, for reports that they are writing inside their own institute. And they're finding that they can use software to manage that whole process much more. What about real specialist jobs like the jobs of doctors? Well this guy here, Vinod Khosla, is a leading VC in Silicon Valley, has said, in his view, that by 10 years time "80% of the functions that doctors do will be done much better and much more cheaply by machines and algorithms.” So there still will be some work for doctors to do, but they probably won't get anything like the same salaries as before. And what's going on here? Let’s step back and look at the technology that’s making all these changes possible. The first is the ability to understand natural language, because in order to be a doctor, you’ve got to understand what people are saying, you need to understand what's written down. And today, if you talk to your phone, sometimes it understands you, and a lot of the time it doesn’t, and it’s a bit of a game. But you know it is getting better and better. And there is this software by IBM which they called Watson, which they developed to understand natural language. They developed it to play in an American quiz show called Jeopardy, which is very popular there, which involves very cryptic clues; it’s like during the worst and most difficult cryptic crossword. And you need an awful lot of background information to be able to jump to the conclusions. Well, they fed Watson the entire works of William Shakespeare, they fed it all of Wikipedia, they fed it the King James Bible, they fed it the iMDB database, they even fed a book that one of these two other people had written. One of them is called Michael Jennings, who is one of the best Jeopardy players ever. He wrote a book called ‘How to win in Jeopardy’ and they let Watson read that too. And it turns out that this Watson did in 2011, have a better reaction in real time. It was disconnected from the Internet, but it was listening to the questions, and it was buzzing in to say “I know the answer to that.” And it got some of them wrong, but it got more of them right than humans. And some people say “Oh, it’s just a game. There’s nothing serious here.” But what IBM is doing next is even more significant. They are using this same software in an advisory role for doctors. They are developing 'Doc' Watson MD, which has the ability to read medical records. And some medical records are in fixed format. But many things in a medical record are just doctors writing down streams of notes. So increasingly Watson can read that. Watson can read all the medical research publications. No human doctor can hope to keep up with all the publications that are written, so Watson's keeping its eye on that. And Watson can clone itself - it can share it own findings, that puts it ahead of the humans as well. It’s able to discern off of new patterns, it’s able to suggest new experiments. So when a patient presents him or herself with various ... symptoms, Watson's sometimes able to say “I'm not sure what's going on here but you could try doing this test now.” And the doctors are able to ask it “Why do you think that?” and Watson’s able to explain its reasoning. And occasionally the doctors say “Oh no, no, you misunderstand this; it can't happen.” And then it learns, and it learns, and all the different Watsons share amongst themselves what they’re learning. And so, it’s not good enough today to replace much of what doctors do, but within 10 years, we’re going to see medical science transformed, and we’re going to see many work that doctors do transformed. Then there’s creative work. And some people say “Oh I don't need to worry about this, you know I'm not a doctor, I’m a creative person. I'm going do very emotionally intelligent stuff. I'm going to be a musician.” Well, there are now computers that can write music in the style of Bach, which other classical music experts can't distinguish - they don't know whether that's been written by a computer, or whether it's a less well known piece by Bach himself. And so on for many other musical superstars. I haven't yet seen the number-one record in the hit parade chart being written by an AI but I don't think it's going to be far away. So there's a bunch of creativity going on there too. So to summarize, what's driving the growth of technological unemployment is a whole series of new technological skills. There’s the understanding of natural language. There's the improvements in reasoning, expert systems, where it’s understanding medical connections and so forth. There’s deep learning which is spotting patterns in data, where Watson and its associated computers have already spotted new ways to make sense out of biopsy data, and for looking at whether somebody's tumor in her breast, whether it’s likely to be malignant or benign. There’s been 5 or 6 tests that have been applied since the 1930s, where Watson and its associations were able to say “look there’s actually another few tests you should be doing,” looking at the tissue alongside as well. So it’s already spotting new patterns. What we’re going to see next is ‘Deep learning ++’ which not only do they spot new patterns, they spot new ways in which they can learn, which is something we humans are still very good at. We are often were confused for a while, then we figure it out. We say “I know how I can learn in this situation! ” Well increasingly, AI is going to do that too. With artificial creativity, with artificial emotion, it’s going to display emotion on its face, which we're going to take as being emotional, and it's also going to be very good, it is already very good in many ways, in detecting our emotions. So there's already some software that you can detect whether people are relaxed or nervous just by looking at their face. Then there's movement. Some people say "Oh we don't need to worry about robot uprising, all we need to do is climb up stairs; robots will never be able to climb up stairs." And today they're not very good at it, but you know what? They’re going to get better. And we can look at some of the videos of Big Dog by Boston Dynamics, now bought by Google, remarkable imagery of robots being able to step around. They're still not very good - the stuff that the robots are trying to do in the Fukushima disaster zone, they’re trying to go in and clean up, they're not able to do that yet - but I think in a few years time they’re going to be better at it. And one reason they're going to be better at it - there’s one missing skill here which is going to make a whole lot of this stuff go much faster, that is the missing skill of computer vision, which means understanding what’s in the environment, because the computers often don't know what's in the environment. The cars - they’re not actually looking at it in the same way that we are. They’ve got lots of kind of radar around or Lidar it’s called. They don't have many normal cameras to figure out what's going on, This is because computer vision still hasn’t been very good. But I do want to spend just a moment telling you about some of the remarkable improvements in computer vision over the last few years. And I’ll refer to a big database. It’s something called the Image-Net database, which computer scientists have assembled and used. It’s got 14 million images, and here’s one, and you can see in here the computer has first figured out there’s a bird there, and then it saw something else in there, it looked and said that’s a frog. And here's another one. It’s looked and said there’s a person, maybe it should have said it's a man, I don't know. And then it said in one hand it's got a hammer and on the other hand it's got a white box, It knows there’s something there, it doesn’t know what it is, so it's left that bit blank, it hasn't been able to figure that out and I'm not sure I could figure it out either; maybe it's a nail of some sort. And it's also spotted a flower pot, which I didn't see in the picture at first, but if you crane your eyes you can see a flower pot, and a power drill. So there are 14 million images like this in the database, and they've been classified by humans, full job if you’ve gone through it very carefully, and they’ve said about a thousand different things are in these pictures, and this is used every year in a big test. You may have heard of very grand challenges like the ‘Grand Challenge DARPA’ used to figure out which self-driving cars were best. Well every year now, there’s a Large Scale Visual Recognition Challenge when they issue ... about a thousand of these pictures, they allow the computers to train on it, and then they give it others and see how accurate or inaccurate they are. And last year, last year the winning software only made 6.6% errors. It was able to recognize correctly, most of the stuff it saw in there. Which was a big improvement from the year before, down from 11.7% errors, to 6.6. In fact over the 4 years this competition’s been running, there’s been a four-fold improvement. Because all these guys go back, they’re working on their algorithms, they’re learning how humans do it. Still not quite as good as humans, because when full humans are asked to do the same thing - not the original humans who did the classification but others are brought in - they generally get (still make some mistakes) they get about ... 95% right, 5% wrong. And soon we might imagine that computers are going to do it better. And late breaking news. In February last year a team for Microsoft Beijing, led by this guy, Jian Sung, they did it outside the normal competition cycle, but they are saying they already got it down to 4.94% error rate. So I think in 10 years time, these systems, they're going to be much cheaper, they’re going to be much smaller, they are going to be telling robots all kinds of things about their environment. And this is key to robots being able to do lots more human tasks. So, time to start summarizing. The challenges ahead. There’s going to be three big challenges ahead. Three big questions [are] probably in your mind by this stage. The first is: Okay, we might be put out of our present job, we might have been a checkout assistant, maybe we can’t do that job anymore. But can’t we retrain, you know? Can’t we go and learn something new? After all, this has happened in the past. After all, 200 years ago in this country, there were many people complaining about the growth of automated weaving. It was the Luddites who complained about it, and said it’s putting them out of jobs. And they managed in the end to get new jobs, maybe not them but their children got new jobs in other kinds of factories. Eventually some of them became software engineers, eventually some of them became stylish hairdressers, and so on. So can’t people retrain? Won’t there be new jobs? And I think yes, today, many humans are able to retrain. But you know what? As robots are getting better and better, as that pace of improvement gets faster, I think that in some relatively near time, perhaps 10 years, perhaps 15 years, perhaps 20 years, most people, if you try and retrain, they take two years out to learn a new skill, they will find out that in the mean time, robots have got better than them. So they thought they were going to do a new job, but robots have got better than them, faster. So that’s why this is going to be different from all the previous times in history. That's why there's more of a threat here. And I think [there] will still be some number of human jobs, but probably not very many. Because some people say “All right, I'm being put out of work by some software, but you know, I can go and become a software writer.” But you don't need so many people writing the software. If you look at say some accounting software by Intuit (Intuit’s a manufacturer of tax preparation software), there used to be maybe, I don’t know, hundreds of thousands of tax accountants around the place. And Intuit came along and wrote their own tax software, and their leaders, the founders of Intuit, became very rich indeed. But you don't need the same hundreds of thousands of people to be doing that. So there’s many fewer jobs available by this winner-takes-all dynamics. So I'm saying here perhaps less than 25%. This guy here, Larry Page, who is the CEO and founder of Google, says that probably maybe 90% of the jobs that people are doing, that there’s really no point doing them soon. He says “Wouldn't the world be a happier place if 90% of the people with jobs put their feet up instead and left the robots to do the work? ” This is in an interview in the Financial Times late last year. It’s worth looking at a few more things he said. He said (same as I’ve just said) that “rapid improvements in artificial intelligence, for instance, will make computers and robots adept at most jobs. Given the chance to give up work, 9 out of 10 people wouldn’t want to be doing what they're doing today.” And would they regret losing their jobs? And he says “This can’t be the right answer.” Just people wanting to get work even if it's inefficient, even if it's just to keep themselves earning, this can't be the right answer. That brings us back to “Well, what is that right answer? ” How can we distribute sufficient income (resources) from the abundance that these robots will be creating? from all the things that they'll be doing? Surely, we should find a way to distribute that to everyone in society. And that brings us to the very last slide. I do think that as well as embracing some of this technological possibility, as well as steering it, we need to work very hard on developing a new social contract. Some people call it UBI - Universal Basic Income - which means that people will get enough to have a good life, regardless of whether or not they are (quotes) "working or not." I don't think we're going to get there straight away. I saw that only one political party made a serious discussion of UBI in the latest election, the Green Party, and they were sort of taken to bits by aggressive interviewing, because they couldn't really explain how they were going to achieve it step by step, and so they stopped talking about it for this election cycle. I don't think it's going to come in overnight, but it can be achieved in stages. Many transitional details need to be worked out. But I say, let's put our brains into working that out. Let's have a grand Apollo-scale project, an Apollo-scale project such as was used to unite lots of effort to put a man on the moon and bring him back in the 1960s. We had an Apollo-scale project in this country in the 1940s, the second half, when we put the National Health Service in place - a grand social reorganization. We need to work on the similar details to transition to something similar for everyone. And it's not just a change in law. It’s a change in zeitgeist, a change in mindset, a change in values that’s needed. Because today when you start discussing this, most people say "You know, we can’t give people money for doing nothing, they’re just going to be lazy. They should be pitied; they should be-... they're undeserving." But I think there’s so much evidence against this, now that we understand more about the brain, now that we understand more about the psychology, there’s lots of evidence that people actually can do great things with the help of ... income support. And look at JK Rowling, who was on a child benefit as she wrote Harry Potter. That may or may not be your idea of the best books ever, but it shows what can be done. So I think more and more of us will be in this situation. And I see finally that the requirement for employment, which many people have dearly wanted to have, many people have long fought for employment, I say this applies only in the initial phase of humanity. I'm looking forward to something called humanity plus. Humanity plus, in which we don't need to work, in which the fruits of robotic work, the fruits of energy from the sun, the fruits of green technology, will be able to be distributed to everybody. The best is ahead, but we’ve got a big social project to work on first. So if we can navigate safely, we'll get there. Change in zeitgeist [is] your part of that project - let's go do it together. Thank you very much. [Applause]

Video Details

Duration: 40 minutes and 48 seconds
Year: 2015
Country: United Kingdom
Language: English
Producer: The Zeitgeist Movement
Director: The Zeitgeist Movement
Views: 116
Posted by: ltiofficial on Jun 18, 2015

David takes the audience through the concept of 'technological unemployment', and how it is poised to greatly affect all of our lives.

Note: This is LTI's 'internal working location' for this video, so please do not publicly pass around this URL. All completed and fully proofread 'official' translations can be found at the Repository location at, which we highly encourage you to embed &/or pass around.

To join/help with these translation efforts: (LTI Forum)

Caption and Translate

    Sign In/Register for Dotsub to translate this video.