1 00:00:06,780 --> 00:00:10,380 Good evening and very warm, welcome to the Oxford Martin School. 2 00:00:10,380 --> 00:00:14,790 Thank you for coming out on this particularly rainy evening. 3 00:00:14,790 --> 00:00:23,010 I think you all in for a treat. For those of you that have just started at Oxford and this is your first term and particularly warm welcome. 4 00:00:23,010 --> 00:00:29,910 The Oxford Martin School is an interdisciplinary group which focuses on the big challenges of the future, 5 00:00:29,910 --> 00:00:35,730 and you'll find a very, very wide range of events going on here on the website. 6 00:00:35,730 --> 00:00:43,140 I was the founding director of the schools in 2006. I came to Oxford in gold and I'm now professor of globalisation, 7 00:00:43,140 --> 00:00:51,960 and one of the enormous pleasures I had was being able to put people together to create very exciting teams. 8 00:00:51,960 --> 00:01:02,520 And so I'm delighted in 2012 to be able to recruit call for a fresh out of his doctorate in Berlin to come and work 9 00:01:02,520 --> 00:01:09,870 on a team we put together on technology and employment the intersection of technological and economic change, 10 00:01:09,870 --> 00:01:16,110 working with Michael Osborne, who was fresh out of his doctorate in the engineering department here. 11 00:01:16,110 --> 00:01:23,850 As you know, the rest is a short history and they have done absolutely fantastic work in 2013, 12 00:01:23,850 --> 00:01:27,870 very soon after the call arrived here and started working with Michael. 13 00:01:27,870 --> 00:01:34,890 They produced a paper which really changed the way that people think about this area. 14 00:01:34,890 --> 00:01:44,940 It's a sign of its success that it has been a source of much controversy and contradiction because it really defined that space. 15 00:01:44,940 --> 00:01:51,420 And many other people have subsequently done work in that area because remains at the head of it. 16 00:01:51,420 --> 00:01:57,900 I stepped down in 2016 and I'm delighted to see that, such as Godfrey, the current director of the school, 17 00:01:57,900 --> 00:02:03,870 is in the back has to leave early, which is why he's not doing what I'm doing. 18 00:02:03,870 --> 00:02:08,850 Karl has gone on to win a number of grants after that. 19 00:02:08,850 --> 00:02:17,640 We've been very fortunate in this work programme to benefit from funding from Citibank, and in fact, they are just renewing that again. 20 00:02:17,640 --> 00:02:24,390 So we absolutely should be able to keep drilling down and pursuing this vital research agenda. 21 00:02:24,390 --> 00:02:28,920 Will we have jobs in the future? What will the jobs be? Where will they be? 22 00:02:28,920 --> 00:02:33,720 And what are the implications in terms of inequality, wages and so on? 23 00:02:33,720 --> 00:02:41,640 The technology trap is a magnificent book. It's for sale afterwards at a massively discounted price of £15. 24 00:02:41,640 --> 00:02:46,020 If you go down the road, you have to buy it for. So buy it here. 25 00:02:46,020 --> 00:02:52,650 And there will also be a drinks reception to which all of you are invited at six o'clock. 26 00:02:52,650 --> 00:03:11,160 So I'm delighted to welcome Karl and for him to launch his new book, The Technology, Trump Capital, Labour and Power in the Age of Automation. 27 00:03:11,160 --> 00:03:18,720 Thank you. It's such a pleasure to be back here to my favourite place to say my favourite lecture hall in the world, 28 00:03:18,720 --> 00:03:27,600 not just because I only have to take the elevator floor down to get here, but also because it's the place where everything actually started. 29 00:03:27,600 --> 00:03:36,870 Back in 2012, when Mike and I got together and began to discussing what the future of the bloc might actually look like, 30 00:03:36,870 --> 00:03:44,340 what how the division of Labour is evolving as artificial intelligence is progressing. 31 00:03:44,340 --> 00:03:48,420 And, as Ian mentioned, has been a source of a bit of controversy. 32 00:03:48,420 --> 00:03:54,060 Hopefully, we'll be able to add to a bit more to that controversy today. 33 00:03:54,060 --> 00:04:03,420 But the theme of the book is very much sort of trying to shed some light on the times we're living in through the lens of history, 34 00:04:03,420 --> 00:04:13,660 because what we estimated back in two thousand twelve is that roughly 47 percent of jobs are at high risk of being automated. 35 00:04:13,660 --> 00:04:18,120 And the question that we've gotten consistently since then is, 36 00:04:18,120 --> 00:04:27,600 is this time different somehow because we've been through periods of rapid technological progress and in the past, 37 00:04:27,600 --> 00:04:36,990 and I found it quite extraordinary when researching this book how much technology has progressed over the past two centuries, 38 00:04:36,990 --> 00:04:43,290 but how little the debates surrounding its effects has actually progressed. 39 00:04:43,290 --> 00:04:55,260 We are now living through an era of automation anxiety again, very similar to what we saw in the 1960s, in the 1930s, in the 1830s and so on. 40 00:04:55,260 --> 00:05:04,410 And it feels quite natural somehow that people shouldn't feel that and to see us think about the tornado that is coming for the office. 41 00:05:04,410 --> 00:05:08,980 And I was actually quite struck to see that a majority of Americans. 42 00:05:08,980 --> 00:05:18,060 This is from a recent Pew Research survey. A majority of Americans now think that there should be restrictions on the number of machines 43 00:05:18,060 --> 00:05:25,770 that businesses should be allowed to implement in order to increase productivity growth. 44 00:05:25,770 --> 00:05:32,760 And this is not that puzzling through the trends of recent automation and debates, 45 00:05:32,760 --> 00:05:40,290 but it is quite puzzling if we look at the long run trajectories of mechanisation and growth. 46 00:05:40,290 --> 00:05:45,990 I studied economic history as an undergrad, one of the first slide you encountered when he studied the economic history. 47 00:05:45,990 --> 00:05:47,190 Is this one right? 48 00:05:47,190 --> 00:05:58,200 It shows that economic growth was roughly stagnant for millennia, a very long time, and took off in extraordinary fashion around eighteen hundreds. 49 00:05:58,200 --> 00:06:05,030 What happens around 18:00 is not that no technology for the first time arrives, 50 00:06:05,030 --> 00:06:13,110 so there was a lot of important inventions before then, like the barometer, the telescope, the horse shoe and so on and so forth. 51 00:06:13,110 --> 00:06:23,130 There's a lot of important inventions happening. But for the very first time, technology translates into higher incomes for the wider population. 52 00:06:23,130 --> 00:06:32,760 And the reason for that must be the mechanised factory, which allowed us to produce more with fewer people before and turned 100 were 53 00:06:32,760 --> 00:06:39,300 few machines that relieved workers of a lot of burdens in their daily jobs. 54 00:06:39,300 --> 00:06:45,990 And as a result of the steady flow of labour saving technologies that have been adopted since then, 55 00:06:45,990 --> 00:06:55,200 we are now able to produce roughly 40 times more in Britain than we were at the dawn of the industrial revolution. 56 00:06:55,200 --> 00:07:02,760 And people's income as a result of that is roughly 40 times higher adjusted for inflation. 57 00:07:02,760 --> 00:07:09,900 Now you might say that that does. Compassion doesn't make any sense because the consumer basket that all of the industrial 58 00:07:09,900 --> 00:07:16,770 revolution was very different and the consumer basket that you combined with your incomes today. 59 00:07:16,770 --> 00:07:20,370 All of these things didn't even exist. 60 00:07:20,370 --> 00:07:30,840 Ordinary people at the time could only look at the lives of the wealthy in the movie, who had servants to do the most tedious things for them. 61 00:07:30,840 --> 00:07:40,290 Today, most of all, so all of us here in this room have access to the electric servant in terms of dishwashers, washing machines, vacuum cleaners. 62 00:07:40,290 --> 00:07:49,830 And so that relieves us of a lot of tedious work, not to mention other inventions like the automobile and out of antibiotics. 63 00:07:49,830 --> 00:07:53,520 And if that wasn't enough evidence of progress, 64 00:07:53,520 --> 00:08:04,350 consider the fact that producing those goods and earning those higher incomes have become so much more comfortable, less well. 65 00:08:04,350 --> 00:08:09,020 Not so long ago, a significant share of the population worked in coal mines. 66 00:08:09,020 --> 00:08:15,020 Evidence and explosions were part of everyday working life, long pieces of often part of the package. 67 00:08:15,020 --> 00:08:19,490 Today, most of us work in air conditioned offices. 68 00:08:19,490 --> 00:08:25,070 The worst thing that happened here at the motor school in recent memory was when the coffee machine broke down. 69 00:08:25,070 --> 00:08:33,980 I think that it puts things in perspective, and if we only look at the sectoral changes and the composition of the workforce, 70 00:08:33,980 --> 00:08:37,110 that also understates the transformation that's taking place. 71 00:08:37,110 --> 00:08:46,250 So if you take the jobs of farm labour in nineteen hundred, that person would have walked the fields with nothing more than animal power. 72 00:08:46,250 --> 00:08:53,840 He would, or she would be exposed to hazardous weather conditions, swarms of insects, all sorts of unpleasant things. 73 00:08:53,840 --> 00:09:03,030 Today, a farm labourer in the industrial west usually sits in his or her tractor and can listen to the music of his or her choice. 74 00:09:03,030 --> 00:09:05,630 So it's quite extraordinary. 75 00:09:05,630 --> 00:09:16,070 What a leap forward that has happened over the past 200 years at the point of the book is that that leap forward is only half of the story. 76 00:09:16,070 --> 00:09:21,530 What did the people actually say when the mechanised factory arrived? 77 00:09:21,530 --> 00:09:26,060 I think there's no doubt that the industrial revolution laid the foundations for modern 78 00:09:26,060 --> 00:09:31,580 growth and much of the increases in growth and well-being that has taken place. 79 00:09:31,580 --> 00:09:36,050 But industrial revolution itself was polarising. 80 00:09:36,050 --> 00:09:42,200 So, for example, Benjamin Disraeli, before he became prime minister of Britain, 81 00:09:42,200 --> 00:09:49,790 wrote a novel cold calling speech in which one character remarks said, I see cities, people with machines. 82 00:09:49,790 --> 00:09:55,460 Suddenly, Manchester must be the most wonderful place of modern times. 83 00:09:55,460 --> 00:10:00,740 The very same year, Friedrich Engels published his conditions of the working classes, 84 00:10:00,740 --> 00:10:09,350 which was written during a stay in precisely Manchester, and Engels had a very different take of what was happening. 85 00:10:09,350 --> 00:10:19,940 He argued that mechanisation only serves to downgrade people as it puts them in the repetitive motions of machinery, which can be deemed unnatural. 86 00:10:19,940 --> 00:10:24,770 And you also argue that it took over a lot of people's jobs, 87 00:10:24,770 --> 00:10:32,750 reduced their wages and incomes to subsistence, and led to the administration of the working class. 88 00:10:32,750 --> 00:10:38,150 We know from history that he was wrong about the future, 89 00:10:38,150 --> 00:10:43,250 but it was actually fairly on target about the period he lived through because what 90 00:10:43,250 --> 00:10:48,590 he witnessed was what the economic historian Bob Ireland has called Engels prowess. 91 00:10:48,590 --> 00:10:55,160 Seven decades of the classic period of the industrial revolution where the British economy first took off, 92 00:10:55,160 --> 00:11:00,800 but very little happened to the wages of ordinary people. 93 00:11:00,800 --> 00:11:03,830 Now the wage data for this period of time isn't great, 94 00:11:03,830 --> 00:11:15,500 but if you look at supplementary sources like based on consumption or biological indicators such as height, they paint a similar picture accords. 95 00:11:15,500 --> 00:11:21,830 Born in 1750, we're actually taller and of course, born in 1850, 96 00:11:21,830 --> 00:11:28,940 suggesting that people's nutrition was adversely impacted as well as their incomes diminished. 97 00:11:28,940 --> 00:11:38,360 And at the heart of the industrialisation process was the replacement of middle income brackets. 98 00:11:38,360 --> 00:11:44,600 So artisan craftsmen who would do sort of every step in the production process themselves, 99 00:11:44,600 --> 00:11:53,570 often in their home, surrounded by their wives and children, their incomes came under pressure as the mechanised factory arrived, 100 00:11:53,570 --> 00:12:01,580 and what the factory did is that it took advantage of cheap labour, child labour in particular. 101 00:12:01,580 --> 00:12:14,080 And it's a really noted fact that the early spinning machines of the industrial revolution were specifically designed to be attended by children. 102 00:12:14,080 --> 00:12:18,830 Manufacturers three that to sap resistance to mechanisation. 103 00:12:18,830 --> 00:12:27,170 They were better off by trying to innovate, innovate in such ways as to get around that. 104 00:12:27,170 --> 00:12:38,990 And children often didn't have much bargaining power, and they only got food allergy, and they were easy to enforce the factory discipline upon. 105 00:12:38,990 --> 00:12:43,520 They were the robots of the industrial revolution, if you like. 106 00:12:43,520 --> 00:12:52,910 And now a great question to economists and economic stories has been why would people 107 00:12:52,910 --> 00:13:00,740 voluntarily have participated in the industrialisation process if it reduced their utility? 108 00:13:00,740 --> 00:13:07,940 The simple answer is that they didn't. They expressed their thoughts and feelings converse like the one you can see above. 109 00:13:07,940 --> 00:13:18,910 They. Son to parliament to block introduction on machinery on several occasions, and they rioted against the Mechanistically Factory, 110 00:13:18,910 --> 00:13:29,440 and all the Luddites achieved was getting the British government to deploy an even larger army against him. 111 00:13:29,440 --> 00:13:36,550 The Army Wellington took against Napoleon in the Peninsula War of 1812 eight 112 00:13:36,550 --> 00:13:41,270 was actually smaller than the army that was sent out against the Luddites. 113 00:13:41,270 --> 00:13:48,610 And I think it's important to remember that a much public commentary focus on the London riots in particular, 114 00:13:48,610 --> 00:13:54,190 was a long wave on machinery riots swept across Britain and even Europe, 115 00:13:54,190 --> 00:13:58,950 and all of them didn't have to do with appalling working conditions of the factory. 116 00:13:58,950 --> 00:14:05,440 You saw similar riots happening against our culture machines as well. 117 00:14:05,440 --> 00:14:15,490 And I think that the economic historian Hobson summarised it nicely when he said that once began with the construction of the 118 00:14:15,490 --> 00:14:24,340 first factories and ended with the construction of the railroads also ended with the publication of the Communist Manifesto. 119 00:14:24,340 --> 00:14:28,330 A lot of these revolutionary technologies became eventually. 120 00:14:28,330 --> 00:14:35,650 The engines of growth also created a lot of political revolutionaries along the way. 121 00:14:35,650 --> 00:14:39,490 The point of the book is not to protect a socialist revolution, 122 00:14:39,490 --> 00:14:49,390 but I think it is noteworthy that levels of income inequality have been increasing over the past three decades or so, 123 00:14:49,390 --> 00:14:56,350 and they are approaching levels not seen since the first industrial revolution. 124 00:14:56,350 --> 00:15:00,880 And there are many factors that are shaping and then income distribution. 125 00:15:00,880 --> 00:15:08,350 There's no question about that. But mechanisation or automation is definitely one of them. 126 00:15:08,350 --> 00:15:13,930 And clearly, it's not driven by textile machinery or steam engines. 127 00:15:13,930 --> 00:15:24,160 This time around, technology is very different. But it is anything having even more pervasive impacts on the labour market. 128 00:15:24,160 --> 00:15:28,690 And as you can see, this process has been going on for some time. 129 00:15:28,690 --> 00:15:37,360 The first electronic computer was invented at the University of Pennsylvania in nineteen forty seven or forty six sorry, 130 00:15:37,360 --> 00:15:43,360 but it consisted of roughly 18000 vacuum tubes, 20 30 tons. 131 00:15:43,360 --> 00:15:48,550 And as a result of that, it didn't have much impact on the labour market. 132 00:15:48,550 --> 00:15:59,080 It took the invention of the microprocessor and later on the invention of the personal computer for businesses to realise that a they 133 00:15:59,080 --> 00:16:09,910 could restructure supply to supply chains in ways that allowed them to take advantage of cheap labour in emerging economies like China. 134 00:16:09,910 --> 00:16:19,090 And secondly, that they could automate routine, repetitive tasks on a mass scale. 135 00:16:19,090 --> 00:16:24,030 And the consequence of that is best shown by this figure. 136 00:16:24,030 --> 00:16:29,590 We see that as productivity growth grows for most of the 20th century. 137 00:16:29,590 --> 00:16:40,180 Wages are growing in tandem. So early machinery is creating a lot of demand for operators whose skills are augmented by machinery, 138 00:16:40,180 --> 00:16:51,430 and the technology complements them in a happy way. That's quite different from robotics, who are replacing those sorts of workers in production. 139 00:16:51,430 --> 00:16:58,850 And as a result of that, you seeing a growing gap between wages and productivity growth. 140 00:16:58,850 --> 00:17:02,900 Now, for all the talk of the of rising inequality, 141 00:17:02,900 --> 00:17:12,320 I think the greatest tragedy is actually that certain groups in the labour markets have been left worse off in absolute terms. 142 00:17:12,320 --> 00:17:18,080 So if you look at the wages of men with no more than a high school degree, 143 00:17:18,080 --> 00:17:24,950 they have actually been falling for prime aged men with no more than a high school figure for four decades now. 144 00:17:24,950 --> 00:17:31,130 So it's not just a question of of the labour market becoming more equal. 145 00:17:31,130 --> 00:17:45,830 Opportunity is diminishing very rapidly for people who used to take on jobs in factories before robots were being adopted on a wider scale. 146 00:17:45,830 --> 00:17:53,270 One of my favourite lines from the book is that if you put one hand in the freezer and the other on the stove, 147 00:17:53,270 --> 00:17:58,940 you should be feeling quite comfortable on average. But we know from experience that that is not the case. 148 00:17:58,940 --> 00:18:04,700 And I think the same can be said about the U.S. labour market, right? 149 00:18:04,700 --> 00:18:08,210 Unemployment is trending around 3.6 percent, right? 150 00:18:08,210 --> 00:18:17,660 It's looking very healthy on the surface. And if you go to certain places, like most coastal cities, it's also doing really well. 151 00:18:17,660 --> 00:18:25,100 But people in the Rust Belt, in particular in cities that specialised in manufacturing industry, 152 00:18:25,100 --> 00:18:32,510 which have suffered from deindustrialisation caused by offshoring and automation, are not doing so well. 153 00:18:32,510 --> 00:18:37,610 You can see if you look at labour force participation rates amongst primates, 154 00:18:37,610 --> 00:18:44,840 men again, you can see that joblessness is on the rise and especially in these places. 155 00:18:44,840 --> 00:18:54,230 There are more robots in Michigan alone, for example, than the entire American West, and that is where most of its problems are. 156 00:18:54,230 --> 00:19:03,980 And and we know from a great body of research that joblessness also comes with a lot of bad social consequences. 157 00:19:03,980 --> 00:19:15,200 Economists tend to think that the purpose of production is consumption, but we actually know from a very big body of research that that's not true. 158 00:19:15,200 --> 00:19:19,340 People attach a lot of meaning to the work. 159 00:19:19,340 --> 00:19:28,340 One of the most consistent findings across countries across different periods of time is that people who work are happier than those who don't. 160 00:19:28,340 --> 00:19:36,470 And if you look at the communities where jobs have dried up, you see that marriage rates are declining. 161 00:19:36,470 --> 00:19:41,540 More children are growing up in single parent households. Crime is on the rise. 162 00:19:41,540 --> 00:19:45,470 Health outcomes are worsening. Suicide and alcohol. 163 00:19:45,470 --> 00:19:52,130 Alcohol and substance abuse is on the rise, and it is not a very healthy development. 164 00:19:52,130 --> 00:19:58,250 And if you want to understand my key three key swing states in Michigan, Wisconsin, 165 00:19:58,250 --> 00:20:12,050 Pennsylvania who had opted for the Democratic candidate in every election since 1992, all of the sudden ended up voting in President Trump in 2016. 166 00:20:12,050 --> 00:20:20,360 Automation is one of the prime reason, and this is a joint research which Shenzhen and taught by the NEC. 167 00:20:20,360 --> 00:20:27,590 And one of the key points is that we have seen nothing yet. We've already seen a backlash against globalisation. 168 00:20:27,590 --> 00:20:34,250 But automation has so far primarily been confined to routine rural based activities. 169 00:20:34,250 --> 00:20:46,760 But the potential scope of what machines can do is expanding very rapidly due to recent advances in machine learning and artificial intelligence, 170 00:20:46,760 --> 00:20:54,620 in particular machines that are increasingly incapable of inferring the rules of the game themselves. 171 00:20:54,620 --> 00:20:58,850 They can tap into the digital trails that we leave behind us. 172 00:20:58,850 --> 00:21:04,820 When we interact online, they can learn through trial and error, and as a result of that, 173 00:21:04,820 --> 00:21:09,650 they're able to perform tasks that are inconceivable to automate. 174 00:21:09,650 --> 00:21:19,820 A few years ago, like driving a truck diagnostic and cease translation work and document review so and so forth. 175 00:21:19,820 --> 00:21:29,630 And and what all of these approaches have in common is that they are driven by increasingly large 176 00:21:29,630 --> 00:21:36,320 datasets that allows machines to predict what a human would have done in any given situation. 177 00:21:36,320 --> 00:21:42,170 So for example, if you want to programme a car to drive in the city of Oxford, 178 00:21:42,170 --> 00:21:50,060 it's almost impossible to foresee every given situation that the vehicle might encounter. 179 00:21:50,060 --> 00:21:58,200 You just can't, and but you can actually fill in the blanks by gathering loads of data. 180 00:21:58,200 --> 00:22:07,620 People driving and then trying to predict what the human driver would have done in that given situation, and I am simplifying a bit, 181 00:22:07,620 --> 00:22:17,610 but it's probably this pattern that is driving everything we're seeing in technology and today and some of you might say, 182 00:22:17,610 --> 00:22:23,910 well, I used Google Translate yesterday, so perfect and autonomous vehicles for all day. 183 00:22:23,910 --> 00:22:35,550 But I think it's important to remember that every technological revolution started with imperfect technology are steam engines, 184 00:22:35,550 --> 00:22:40,770 for example, were mainly used to train coal mines. Yet they were the ones in. 185 00:22:40,770 --> 00:22:50,310 So that eventually became the prime movers of the industrial revolution and powered economic growth for a very long time. 186 00:22:50,310 --> 00:22:59,790 And I think that we are actually very much underestimating the potential impacts that I could have on labour markets. 187 00:22:59,790 --> 00:23:07,980 The first reason being that machines don't have to be perfect in order to outperform us because we certainly are not. 188 00:23:07,980 --> 00:23:12,420 So this is a study of Israeli judges and their decision making during the day. 189 00:23:12,420 --> 00:23:16,380 And you can see that early in the morning after we had a morning snack, 190 00:23:16,380 --> 00:23:24,840 insulin levels are high and we make a very high share of favourable positions that then tends to drop off during the morning, 191 00:23:24,840 --> 00:23:29,070 bumps out after breakfast again and the same after lunch. 192 00:23:29,070 --> 00:23:36,420 And we are very much shaped by what we eat, how much we sleep. 193 00:23:36,420 --> 00:23:47,760 Some of us have bad temper in certain situations. So machines are actually have a comparative advantage in a wide range of tasks just because of that. 194 00:23:47,760 --> 00:24:02,040 The second point is that for automation to happen, machinery doesn't necessarily have to replicate every motion that the human does in his or her job. 195 00:24:02,040 --> 00:24:09,990 So we didn't automate away the jobs of lamplight as, for example, by building robots capable of climbing lampposts. 196 00:24:09,990 --> 00:24:16,890 We didn't automate away the jobs of known dresses by building robots that can walk out of the home. 197 00:24:16,890 --> 00:24:25,440 Shop trees carry wood and piles of water into them, and heat the water on the stove and then perform the motions of hand-washing, right? 198 00:24:25,440 --> 00:24:32,400 We did that by inventing the electric washing machine that does an entirely different set of motions. 199 00:24:32,400 --> 00:24:40,080 And because engineers are very good at finding these clever ways of restructuring tasks to make them automatable, 200 00:24:40,080 --> 00:24:45,750 potential scope of automation is much greater than many people and actually think. 201 00:24:45,750 --> 00:24:54,600 Nonetheless, there are certain domains in which human workers still hold the comparative advantage. 202 00:24:54,600 --> 00:25:01,230 And this is joint research with Mike Osborne at the Wharton School that goes back to 2013. 203 00:25:01,230 --> 00:25:09,390 And what we did is we tried to sort of think about in which domains machines still perform very poorly, 204 00:25:09,390 --> 00:25:14,370 where there's been very little progress in recent years. 205 00:25:14,370 --> 00:25:20,460 And I think what example that illustrates this quite well is Turing. 206 00:25:20,460 --> 00:25:27,000 Test competitions for efficient bots try to convince human judges of them being human. 207 00:25:27,000 --> 00:25:34,110 And a lot of pundits argued, I think was about three years ago now that there was a big breakthrough because 208 00:25:34,110 --> 00:25:41,880 one chat bot managed to convince 30 percent of human judges of its being a person. 209 00:25:41,880 --> 00:25:53,220 But it did so by pretending to be a 13 year old orphan Russian boy speaking English as a second language with no understanding of English culture. 210 00:25:53,220 --> 00:26:00,300 And if you think about the variety of much more complex social interactions that you do in your daily jobs, 211 00:26:00,300 --> 00:26:06,150 we try to persuade people that you right to negotiate certain things. 212 00:26:06,150 --> 00:26:17,040 You try to motivate your colleagues. It's almost inconceivable we will have an algorithm that's able to perform that in the foreseeable future. 213 00:26:17,040 --> 00:26:23,160 And I think the same is true of creativity, and there's a big debate in the machine learning community, 214 00:26:23,160 --> 00:26:29,710 admittedly, with regard to whether algorithms can be creative. 215 00:26:29,710 --> 00:26:36,060 But I think people who who think that they can be or things that they are tend to conflate 216 00:26:36,060 --> 00:26:42,360 novelty and creativity so I can draw something here on the world and call myself an artist. 217 00:26:42,360 --> 00:26:47,430 A few of you would be likely to buy my paintings. 218 00:26:47,430 --> 00:26:56,340 And the tricky part here is not to generate something that's novel is generating something that's novel and makes sense. 219 00:26:56,340 --> 00:27:06,110 So if you think about. For example, software writing, classical music, you can certainly take your favourite symphonies, 220 00:27:06,110 --> 00:27:11,370 so maybe they'll know that and say that these are the best symphonies that have ever been written. 221 00:27:11,370 --> 00:27:21,720 And then, you know, allow the algorithm to come up with some combination of that, and it's very likely to sound quite similar to Mozart. 222 00:27:21,720 --> 00:27:25,830 You're not going to arrive with Stravinsky or Schoenberg by doing that. 223 00:27:25,830 --> 00:27:33,990 And I think one of the reasons for that is that when humans are creative, we draw upon based on experience from all walks of life. 224 00:27:33,990 --> 00:27:45,690 Sometimes even a dream. And in many cases, when task create required creativity, much of the data is likely to be outside with the training dataset. 225 00:27:45,690 --> 00:27:52,320 Last bottlenecks may be the least intuitive one that relates to the perception of manipulation of irregular objects. 226 00:27:52,320 --> 00:27:54,750 So some very easy things for us, 227 00:27:54,750 --> 00:28:03,630 distinguishing between a piece of rubbish that's on the floor or a really important document straightforward for most of us to do. 228 00:28:03,630 --> 00:28:06,840 It's actually so easy to explain similar tasks like, you know, 229 00:28:06,840 --> 00:28:18,090 distinguishing between a pot that is stuck in holds and they start in the pot that tells the plant also quite intuitive tasks not so easy to automate. 230 00:28:18,090 --> 00:28:26,220 So unfortunately, I think the job of the cleaner is one of the last we are likely to see is appearing. 231 00:28:26,220 --> 00:28:35,550 And that being said, there are a lot of jobs that are not very intensive and tasks does require creativity and complex social interactions. 232 00:28:35,550 --> 00:28:43,380 And if in fact, the majority of jobs in transportation, logistics, retail, construction don't. 233 00:28:43,380 --> 00:28:49,950 And as a result of that, the potential scope of automation is quite significant. 234 00:28:49,950 --> 00:28:56,460 So you can think about three 3.5, five million cashiers that are employed in the US today. 235 00:28:56,460 --> 00:29:00,090 If it goes into the Amazon Ghost story, we won't see a single one. 236 00:29:00,090 --> 00:29:05,280 There's another 3.5 million truck, taxi and bus drivers in autonomous vehicles. 237 00:29:05,280 --> 00:29:12,750 A right? All of them more likely to be exposed to those technological developments. 238 00:29:12,750 --> 00:29:19,600 And there's another 2.2 million people in the US still working in call centres, just to name a few examples. 239 00:29:19,600 --> 00:29:25,470 And when we published this study a few years ago now, 240 00:29:25,470 --> 00:29:35,910 we actually also published a list of 32 list of some 700 to occupations and their relative exposure to automation. 241 00:29:35,910 --> 00:29:42,210 And you can imagine that a lot of people pick it up from some of those and said, Well, this is silly, doesn't make sense to us. 242 00:29:42,210 --> 00:29:51,840 It from my friend Ken Cukier, the economist, used to tease us because we found that fashion models are highly exposed to automation. 243 00:29:51,840 --> 00:29:54,660 These fashion models here actually don't exist. 244 00:29:54,660 --> 00:30:05,520 They have been generated through generative adversarial networks using thousands of pictures, and they're already being used by companies like yours. 245 00:30:05,520 --> 00:30:11,910 And and I think if we only look at the potential scope of automation, 246 00:30:11,910 --> 00:30:19,920 we actually also miss a lot of stuff because we're only looking at the first order effects. 247 00:30:19,920 --> 00:30:25,110 So a lot of second and third order effects as well that are shaping the cloud. 248 00:30:25,110 --> 00:30:28,950 So in the early days of electrification, for example, 249 00:30:28,950 --> 00:30:39,690 all we did was replacing the steam engine with an electric motor at a sense as the central power source of the factory. 250 00:30:39,690 --> 00:30:43,980 And that didn't have much of an impact on productivity growth, right? 251 00:30:43,980 --> 00:30:46,800 Because everything else remained intact. 252 00:30:46,800 --> 00:30:56,920 All of these shafts and counter shafts in the factory were still there, but it took a while for engineers to figure out actually what you can do. 253 00:30:56,920 --> 00:31:07,290 You can place an electric motor on every single machine and you can then sequence them in the natural flow of production. 254 00:31:07,290 --> 00:31:11,070 And that's the approach that gave rise to mass production. 255 00:31:11,070 --> 00:31:20,160 It is what allowed Henry Ford to produce the Ti model at a sufficiently low price for it to become the people's vehicle. 256 00:31:20,160 --> 00:31:28,500 And the same approach to doing things spread gradually across industries. 257 00:31:28,500 --> 00:31:35,490 And that was when the main effects on the economy and productivity growth was felt. 258 00:31:35,490 --> 00:31:46,930 If you look at some of the early horse carriages or the early automotive on such as, it actually almost looked like a horse carriage, right? 259 00:31:46,930 --> 00:31:58,180 So we're done this essentially replacing the horse with an internal combustion engine as the prime mover of your vehicle. 260 00:31:58,180 --> 00:32:03,250 And it took a while for us to figure out that, well, first of all, we can redesign the entire car. 261 00:32:03,250 --> 00:32:05,710 It doesn't have to look like a horse carriage. 262 00:32:05,710 --> 00:32:12,640 And we then realised that religion is a lot of gasoline stations and we need to build in a lot of shops along the roads, 263 00:32:12,640 --> 00:32:17,430 which gave rise to road commerce, and more tourists have created millions of jobs. 264 00:32:17,430 --> 00:32:22,780 It took a while for us to figure out that we need to build interstate highways, 265 00:32:22,780 --> 00:32:27,310 which gave rise to the trucking industry and changed distribution models. 266 00:32:27,310 --> 00:32:31,840 And it took a while for people to figure out that they actually don't have to live in 267 00:32:31,840 --> 00:32:36,790 congested cities and the longer they could live in the suburbs and commute into work. 268 00:32:36,790 --> 00:32:41,020 So there was the mobile from the early days just replace the horse. 269 00:32:41,020 --> 00:32:47,140 But over a longer period of time, it reshaped the structure of the entire economy. 270 00:32:47,140 --> 00:32:54,010 And I think we can expect the same from artificial intelligence and autonomous vehicles. 271 00:32:54,010 --> 00:32:56,260 I'm an economist, I'm not a futurist. 272 00:32:56,260 --> 00:33:05,260 I will do this very briefly, but there are certain things that I think we can actually infer when autonomous vehicles emerge. 273 00:33:05,260 --> 00:33:13,180 So first of all, you're very unlikely to need a lot of parking space in cities on the sidewalks so you can measure roads slim 274 00:33:13,180 --> 00:33:21,760 build a parking space success because the costs actually just going to satellite a satellite out of the cities. 275 00:33:21,760 --> 00:33:28,180 And you can imagine that if everybody has their own automated chauffeur and they're also more 276 00:33:28,180 --> 00:33:34,660 likely to live farther away from work and do some of the work in the cars on the way home. 277 00:33:34,660 --> 00:33:44,740 So density is maybe not going to be that much of an issue in city centres and a longer as a result of that. 278 00:33:44,740 --> 00:33:49,210 And lastly, the car itself is clearly going to look very different. 279 00:33:49,210 --> 00:33:54,880 Traffic may not look as this 1950s version with a family playing board games. 280 00:33:54,880 --> 00:34:02,530 You may have a Netflix subscription instead, or the car may have a minibar, and that will be the most expensive part of your trip. 281 00:34:02,530 --> 00:34:14,850 But what is clear is that it has the potential to change a lot in terms of how we live and work, provided that we let it do that. 282 00:34:14,850 --> 00:34:25,810 And because I think that the economy's long tiff us out to something when it suggested that what's happened on the farms, 283 00:34:25,810 --> 00:34:34,270 if if horses could have joined the Democratic Party voted, what happened on the farms might have turned out differently. 284 00:34:34,270 --> 00:34:36,490 They could have used their political clout, 285 00:34:36,490 --> 00:34:46,930 a political voice to bring the spread of the tractor to a halt and going already seeing some examples of that. 286 00:34:46,930 --> 00:34:51,460 So this is truck drivers who went on strike in the state of Missouri the other week, 287 00:34:51,460 --> 00:34:59,020 demanding legislation to block the introduction of autonomous trucks as they are getting more pervasive. 288 00:34:59,020 --> 00:35:09,370 This is dockworkers who went on strike in Los Angeles Harbour a couple of months ago, fearing autonomous cargo trucks. 289 00:35:09,370 --> 00:35:17,740 They did misunderstand on paper slap. It's not. It's not the conspiracy not planning to automate this 47 percent of the jobs. 290 00:35:17,740 --> 00:35:27,910 But what it shows is that the societal response to automation is to some extent, already taking place. 291 00:35:27,910 --> 00:35:34,360 When I began to write the book a couple of years ago, robotaxis wasn't on anybody's radar. 292 00:35:34,360 --> 00:35:41,020 Now it's being discussed on both sides of the Atlantic. Bill de Blasio proposed a blue robot tax the other week, 293 00:35:41,020 --> 00:35:49,240 and worker councils would essentially decide whether businesses should be allowed to automate or not and whether it's essential for the company. 294 00:35:49,240 --> 00:35:56,350 So I think we're seeing to some extent some of the backlash against automation already happening. 295 00:35:56,350 --> 00:36:01,780 The point is not that there won't be jobs in the future. 296 00:36:01,780 --> 00:36:04,540 I'm absolutely certain that there will be. 297 00:36:04,540 --> 00:36:11,230 And if you go back to nineteen hundred right and you are somebody of your grandmother or a great grandfather, 298 00:36:11,230 --> 00:36:16,660 uh, what do you think will be the jobs of the early 20th century? 299 00:36:16,660 --> 00:36:18,010 They wouldn't have sat well. 300 00:36:18,010 --> 00:36:27,320 My daughter is certainly going to be a software engineer and my son is going to be hot yoga teacher, and that would be very unlikely. 301 00:36:27,320 --> 00:36:34,150 And I think in a similar fashion, you are very much sort of struggling to predict the jobs that will emerge in the future. 302 00:36:34,150 --> 00:36:41,470 We can't at least now cost some things that are happening in the labour market right now. 303 00:36:41,470 --> 00:36:48,100 Lincoln did a little survey on their platform a while ago, looking at new and emerging jobs. 304 00:36:48,100 --> 00:36:55,390 And amongst those, you can find arrest developers and data scientists and social media interns and zoom by instructors. 305 00:36:55,390 --> 00:37:02,920 And you can also find big data architects to. For marketing specialists and the beach body coach, I won't disappoint with this one. 306 00:37:02,920 --> 00:37:10,840 I think this actually reflects a broader pattern that we've seen in the labour market over the past couple of decades. 307 00:37:10,840 --> 00:37:18,460 We see that new jobs are emerging in very skilled industries that are highly clustered, 308 00:37:18,460 --> 00:37:27,830 and those people are earning higher wages, are demanding a lot of in-person type of services that are hard to automate. 309 00:37:27,830 --> 00:37:34,090 And you can see this even in official statistics now. So some of the air, for example, that's existed for a long time. 310 00:37:34,090 --> 00:37:41,230 It's only very, very recently became so job title because we tend to demand more of them as we grow 311 00:37:41,230 --> 00:37:46,750 wealthier and this is often joined by other targets or better a couple of years ago. 312 00:37:46,750 --> 00:37:54,580 And what it shows is that there's been a shift in where these new jobs are actually emerging. 313 00:37:54,580 --> 00:38:02,710 So before the computer revolution of the 1980s, new jobs were actually recently dispersed across space. 314 00:38:02,710 --> 00:38:09,820 Since then, and you can also see in the statistics of more of these job titles directly relate to computer technologies. 315 00:38:09,820 --> 00:38:14,890 They have become increasingly clustered in skill cities. 316 00:38:14,890 --> 00:38:23,080 And what happens when you create, when you take a job in a city like London or Oxford is that that person goes out to the local economy, 317 00:38:23,080 --> 00:38:29,530 goes to the hairdresser, goes grocery shopping and uses the use, the transportation system and so on and so forth. 318 00:38:29,530 --> 00:38:36,850 And that creates an average five new jobs in the local non-traded service economy. 319 00:38:36,850 --> 00:38:43,330 As a consequence of that, economic activity has become much more clustered. 320 00:38:43,330 --> 00:38:47,830 Some of you may remember this sort of debate and the end of the 1990s, 321 00:38:47,830 --> 00:38:53,260 where people predicted that of the office and Tom Friedman famously wrote this book The World is flat. 322 00:38:53,260 --> 00:39:01,540 But since that the opposite has actually happened. Economic activity has become increasingly clustered in these skilled cities. 323 00:39:01,540 --> 00:39:11,530 And the flip side of that is that where robot technologies or automation technologies have been adopted, the exact opposite has happened. 324 00:39:11,530 --> 00:39:21,760 So in other manufacturing cities where jobs have been automated away, the local service economy has taken a hit as well. 325 00:39:21,760 --> 00:39:28,120 And as a result of that, we see many of these social problems that I mentioned earlier. 326 00:39:28,120 --> 00:39:37,180 Now, so in conclusion, what I think is the key message of this book is that in a way we have been here before, 327 00:39:37,180 --> 00:39:46,270 and in a way resistance to technological change has been the historical norm rather than the exception. 328 00:39:46,270 --> 00:39:55,900 And I think as a great economist, Joseph Schumpeter observed that technological progress also involves creative destruction in employment, 329 00:39:55,900 --> 00:40:07,120 meaning that there will be both winners and losers. And if a lot of people see their jobs and incomes under threat, they are likely to act against it. 330 00:40:07,120 --> 00:40:16,030 And one reason that economic growth was actually slow for so long is that workers vehemently 331 00:40:16,030 --> 00:40:21,730 resisted any technology that threatened their jobs and incomes and fearing social unrest. 332 00:40:21,730 --> 00:40:25,870 Monarchs and governments typically sided with their pockets. 333 00:40:25,870 --> 00:40:31,150 And I think what we need to do is that we need to make sure that people benefit from 334 00:40:31,150 --> 00:40:38,590 these technologies also in the short run to ensure continued acceptance for them. 335 00:40:38,590 --> 00:40:43,270 Because in the end of the day, technological progress isn't an actual force. 336 00:40:43,270 --> 00:40:50,320 It requires societal acceptance. And the great difference to the industrial revolution this time around is that 337 00:40:50,320 --> 00:40:55,300 we are in a much better position to actually manage and shape the outcome. 338 00:40:55,300 --> 00:41:01,600 And so the book concludes with a few policy proposals, which I won't go into here. 339 00:41:01,600 --> 00:41:04,750 Say, If you want to read about those who actually have to buy the book, 340 00:41:04,750 --> 00:41:09,070 it's a great gift for family and friends, and maybe you can open for discussion. 341 00:41:09,070 --> 00:41:23,620 Thanks very much. Thanks very much for that. 342 00:41:23,620 --> 00:41:33,310 Very, very fascinating. Short summary of what is immensely readable and important book. 343 00:41:33,310 --> 00:41:45,640 We have about 15 minutes. This is being webcast and video, so be aware that if you ask a question that you might be broadcast live around the world. 344 00:41:45,640 --> 00:41:49,420 And with that, I hope not intimidating comment. 345 00:41:49,420 --> 00:41:55,030 Who'd like to pose a question to Carl? Thank you. 346 00:41:55,030 --> 00:42:05,560 Can you wait for the microphone, please? Thank you. 347 00:42:05,560 --> 00:42:13,510 My name's young, fun partner of an investment firm investing in education businesses, in part to address some of these challenges, you've pointed out. 348 00:42:13,510 --> 00:42:21,370 Carl, if you could wave a magic wand to ensure that as many people as possible benefit from these 349 00:42:21,370 --> 00:42:28,900 new technologies as opposed to being pushed down into underemployment or low skilled jobs? 350 00:42:28,900 --> 00:42:37,840 What would you what would you like to do? Well, if I answer your question, you're basically giving away the selling proposition of the book. 351 00:42:37,840 --> 00:42:39,910 But I'll give it a go. 352 00:42:39,910 --> 00:42:46,630 So I think unfortunately, there's a tendency to say that we have this really big challenge and we need one big solution to solve it. 353 00:42:46,630 --> 00:42:48,880 And that's universal basic income. 354 00:42:48,880 --> 00:42:58,870 I think there is a lot of things that can be done without my massive mind individually that can make a big difference collectively. 355 00:42:58,870 --> 00:43:02,410 And as I mentioned a couple of times during the presentation, 356 00:43:02,410 --> 00:43:11,710 I think much actually lies in the economic geography of new jobs and the fact that we're seeing this sort of divergence between metropolitan cities, 357 00:43:11,710 --> 00:43:16,930 the countryside and old manufacturing industries. And I think that is really sort of tearing apart society. 358 00:43:16,930 --> 00:43:23,920 And so one example which which I which is close to my heart because it's close to where I grew up in southern Sweden, 359 00:43:23,920 --> 00:43:27,430 is the city of Melbourne, which had specialised in building ships. 360 00:43:27,430 --> 00:43:37,180 And in the early 1990s, Cox cocoons. The shipyard closed down and the city was actually doing very poorly for a long time, 361 00:43:37,180 --> 00:43:41,920 and its revival came with the construction of the Harrison Bridge to Copenhagen. 362 00:43:41,920 --> 00:43:52,060 And so all of the sudden people in management statements live where housing is cheap tap into booming labour markets. 363 00:43:52,060 --> 00:44:01,000 Most of them would spend their earnings locally where they live, which gave a boost to local service economy and created a virtuous cycle. 364 00:44:01,000 --> 00:44:05,580 And it's actually one of the most dynamic labour markets in Europe. 365 00:44:05,580 --> 00:44:13,690 And so I think a lot can be done by actually connecting, declining and expanding real regions. 366 00:44:13,690 --> 00:44:22,940 There's another I think there's 12 in total policy proposals I got through, so I leave the 11 remaining ones as to selling proposition. 367 00:44:22,940 --> 00:44:30,140 And thank you for that question. Yeah. Um, it's a very left. 368 00:44:30,140 --> 00:44:38,500 It's sort of left hand. Thanks for a brain talk, really enjoying it. 369 00:44:38,500 --> 00:44:40,960 I was whenever these conversations come up, 370 00:44:40,960 --> 00:44:49,150 I think about the artisans who were threatened in the industrial revolution with the loss of their livelihoods. 371 00:44:49,150 --> 00:44:55,840 And yet I would imagine in the room today and in Oxford, particularly in a very high, highly affluent place. 372 00:44:55,840 --> 00:45:02,300 Generally, there are a lot of people wearing artisan wristwatches and eating artisan chocolate 373 00:45:02,300 --> 00:45:09,130 in something of a rebuttal of the idea that mass produced goods are best. 374 00:45:09,130 --> 00:45:19,900 Do you think that there is some kind of future for things still made by hand or things made in a bespoke way that as people get wealthier and living, 375 00:45:19,900 --> 00:45:26,950 standards become higher? There may be some greater resistance to the idea of mass-produced things or not 376 00:45:26,950 --> 00:45:30,640 solutions coming from large datasets as opposed to individual tailored ones. 377 00:45:30,640 --> 00:45:34,760 Thank you. And it's a great question, so I think, first of all, 378 00:45:34,760 --> 00:45:40,760 it's important to note that it's still a very small part of the overall economy, but it is growing. 379 00:45:40,760 --> 00:45:48,890 And I think the last part you mentioned is absolutely key, which is that you're saying that as we grow richer, we may demand more of that. 380 00:45:48,890 --> 00:45:51,230 And I think that's exactly what's happening. 381 00:45:51,230 --> 00:45:57,800 But that also means it's not going to be a relief to many of the places that are seeing this adverse consequences of automation. 382 00:45:57,800 --> 00:46:05,900 Because if you're producing a T-shirt that is selling for 40 euros, you need to be have some people around that are likely to buy them. 383 00:46:05,900 --> 00:46:15,290 And one city that the city of Ballarat, reason that the city of Berlin is a very attractive place to live but is essentially bankrupt, 384 00:46:15,290 --> 00:46:22,560 is that, you know, to produce many of these wonderful things, you actually need consumers well in the long run. 385 00:46:22,560 --> 00:46:31,900 Otherwise it doesn't really pay. So have. 386 00:46:31,900 --> 00:46:40,630 Thank you. You've talked about it almost as a process that has happened since the beginning of capitalism and industrialisation, 387 00:46:40,630 --> 00:46:48,310 the substitution of debt, capital for life, labour. 388 00:46:48,310 --> 00:46:56,500 What opportunities to this revolution in the way we produce and distribute goods? 389 00:46:56,500 --> 00:47:00,100 Is it similar to whether you accept it or not? 390 00:47:00,100 --> 00:47:06,370 A current strategy of cycle with major structural opportunities for new employment, 391 00:47:06,370 --> 00:47:13,540 for new activity, and particularly with reference to climate change? 392 00:47:13,540 --> 00:47:21,640 Yeah, absolutely. I think that's right. I mean, one of the fastest growing occupations today is that while also solar panel installers, right? 393 00:47:21,640 --> 00:47:24,130 So there is massive opportunity there. 394 00:47:24,130 --> 00:47:33,910 And I think Faultline and their new trillion dollar invest ment initiative in Europe to spur not just job creation, 395 00:47:33,910 --> 00:47:39,970 but, you know, combat climate change. And I do think that that is a sector that's likely to expand. 396 00:47:39,970 --> 00:47:40,960 That said, though, I mean, 397 00:47:40,960 --> 00:47:48,850 many of these new industries don't provide the same sort of broad based opportunity as the mass production industries of the 20th century did. 398 00:47:48,850 --> 00:47:51,940 But they tend to be high income industries. 399 00:47:51,940 --> 00:47:59,240 They support a lot of the incomes of people that provide in-person Typekit services, so the multiplier is quite high. 400 00:47:59,240 --> 00:48:05,590 But broadly speaking, yes, I think that's right. Hi, thank you for the talk. 401 00:48:05,590 --> 00:48:11,230 So my question is, why do you reckon will be the effect of automation on the military, 402 00:48:11,230 --> 00:48:15,760 such as, for example, I intend to develop and mass produce androids, human eyes? 403 00:48:15,760 --> 00:48:23,480 While Iraq won't be the effect on the, let's say, the U.S. military and on first generation where warfare. 404 00:48:23,480 --> 00:48:27,020 Well, it's a great place, and it's not one I spent much time on, to be honest with you, 405 00:48:27,020 --> 00:48:36,360 I think we're living in a very good report and that's for the last couple of years back and with some research looking into that. 406 00:48:36,360 --> 00:48:44,150 And clearly, there are huge incentives to adopt automation technologies in the military where the cost of lives is very high. 407 00:48:44,150 --> 00:48:52,280 And we know from the past that the military has also been a key driver of innovation and in many sectors. 408 00:48:52,280 --> 00:49:00,030 So I assume the role of the military is quite significant, but it's nothing I've looked into possible. 409 00:49:00,030 --> 00:49:04,540 Wants to see a show of hands. All right. 410 00:49:04,540 --> 00:49:10,510 And women are asking questions, does the lady justice? All right. Give us a major step. 411 00:49:10,510 --> 00:49:14,410 Yeah, you're right. Yes. 412 00:49:14,410 --> 00:49:18,610 Thank you. Thank you for the fascinating talk. 413 00:49:18,610 --> 00:49:30,820 I'm 20 percent into your book on my Kindle. I think technology is the end to the mean, and we are talking about the end, not the mean. 414 00:49:30,820 --> 00:49:37,360 We, the humans are responsible for technological progress and we know and we cannot deny the 415 00:49:37,360 --> 00:49:43,030 fact that technology will evolve and you will progress and it will continuously progress. 416 00:49:43,030 --> 00:49:49,990 And we are looking at the future of technology, which means I haven't got into your policy section yet. 417 00:49:49,990 --> 00:50:02,920 Maybe you have made this recommendation. So what type of policy and what could governments do to look at the future and predict the future and prepare 418 00:50:02,920 --> 00:50:12,340 the human race to face this technological progress and so that we are not trapped in this technological trap, 419 00:50:12,340 --> 00:50:17,810 as you call it? So that's my question. Thank you. Well, that's an easy one. 420 00:50:17,810 --> 00:50:22,700 And so I think one of the key points of the book is actually that, you know, 421 00:50:22,700 --> 00:50:32,360 technological progress is far from inevitable and if it was industrial, revolution would have happened a bit earlier in the history of mankind. 422 00:50:32,360 --> 00:50:41,870 If it was, every country would have adopted the same technologies to the same extent and every country would be rich as a consequence of that. 423 00:50:41,870 --> 00:50:45,530 But that's just not what we see happening over the past 200 years. 424 00:50:45,530 --> 00:50:53,930 There's actually been income divergence rather than convergence, which is a bit of a challenge there. 425 00:50:53,930 --> 00:51:07,970 And so I don't think that it's, you know, easy to predict future economic outcomes on the basis of what is happening in technology alone, 426 00:51:07,970 --> 00:51:15,470 just because the existence of mere existence of technology doesn't mean that it's going to be widely adopted. 427 00:51:15,470 --> 00:51:21,490 So I think you need to look at different factors in how they interact, and technology is one of them. 428 00:51:21,490 --> 00:51:29,880 The distribution of political power in society is one of the reason that mechanisation was adopted a mass scale in Britain. 429 00:51:29,880 --> 00:51:33,470 The first was that the merchant class that grew well, 430 00:51:33,470 --> 00:51:39,890 variables became more politically influential and they would have nothing that jeopardised 431 00:51:39,890 --> 00:51:44,030 their incomes and mechanisation was deemed essential to British competitiveness. 432 00:51:44,030 --> 00:51:47,460 And as in trade and others, that's increasingly integrated. 433 00:51:47,460 --> 00:51:57,280 Markets meant that the political power of the craft skills was significantly eroded because it didn't extend beyond their own cities. 434 00:51:57,280 --> 00:52:00,410 So you saw this sort of great shift in the balance of political power, 435 00:52:00,410 --> 00:52:05,270 and that's the sort of one of the the reason that it's first happened in Britain. 436 00:52:05,270 --> 00:52:13,280 So I think we need to, you know, look at all of this as more of a complex system we have within our natural part of modern schools, 437 00:52:13,280 --> 00:52:20,680 a group that is doing great work in that domain. And I would encourage you to look at the website. 438 00:52:20,680 --> 00:52:27,770 That's never the person that I think it will be when we have time for. 439 00:52:27,770 --> 00:52:36,800 So you are typically on like a hot topic about like attacks on robots and something that I'm concerned about, 440 00:52:36,800 --> 00:52:42,030 that we already see the gap between the frontier of the productivity. 441 00:52:42,030 --> 00:52:49,780 The firms on the frontier of the productivity and the firms are trying to catch up is increasing. 442 00:52:49,780 --> 00:52:53,450 Uh, what do you think about this tax? 443 00:52:53,450 --> 00:53:02,890 I mean, are you worried that this kind of tax maybe actually even make it harder for firms that are trying to catch up to the frontier of technology? 444 00:53:02,890 --> 00:53:05,030 Um, it would be more expensive. 445 00:53:05,030 --> 00:53:15,620 It would be harder for them and how we could come up with this smart kind of taxation that take into account this problem. 446 00:53:15,620 --> 00:53:19,580 So that's a great question. I think I mean, in a way, almost to answer it yourself. 447 00:53:19,580 --> 00:53:24,710 Obviously, if you're a pilot about the robotaxis becomes more expensive to adopt these technologies. 448 00:53:24,710 --> 00:53:28,320 If you're behind the frontier, it gets more expensive to catch up. 449 00:53:28,320 --> 00:53:36,530 So there you have it. I mean, I don't really understand why we should single out robots as a single source of taxation. 450 00:53:36,530 --> 00:53:41,900 I think that's a driver of productivity growth, things we want to see more office, not the things you want to tax. 451 00:53:41,900 --> 00:53:46,190 I think we need to think about taxing capital more broadly. 452 00:53:46,190 --> 00:53:51,920 Clearly, if you see the labour share of income falling over four consecutive decades, you would think that's, you know, 453 00:53:51,920 --> 00:53:58,790 that is going to have certain consequences for public finances, which has being worked on at my school as well. 454 00:53:58,790 --> 00:54:10,970 I'm pleased to say I think maybe even by your group and and and and so, so, so clearly that is a very, very grave concern. 455 00:54:10,970 --> 00:54:16,220 And I think we need to rebalance sort of at tax rates between capital and labour and going 456 00:54:16,220 --> 00:54:21,460 forward also took place close many of the loopholes and loopholes that currently exist. 457 00:54:21,460 --> 00:54:22,850 Great last person there. 458 00:54:22,850 --> 00:54:31,510 And then Carl will be signing books and he can also speak to him over drinks, which will be even more pleasant than being an elected official. 459 00:54:31,510 --> 00:54:34,360 OK, so my question is, 460 00:54:34,360 --> 00:54:42,820 is that you said we've gone through we've gone through the cycle of new technologies coming in and people being very scared but before. 461 00:54:42,820 --> 00:54:48,940 So my question is what is one lesson that we're able to draw from the past in terms 462 00:54:48,940 --> 00:54:53,440 of actually learning how to deal with these fears that we can apply to today? 463 00:54:53,440 --> 00:54:58,660 So I think one of the key lessons you see from the first industrial revolution is that the economic ideas matter, right? 464 00:54:58,660 --> 00:55:02,650 People at the time believe that the world was little too right. 465 00:55:02,650 --> 00:55:09,110 Larger incomes without the resources in larger populations with no sort of increase in per capita terms. 466 00:55:09,110 --> 00:55:15,070 As a result of that, any redistribution of income was deemed counterproductive. 467 00:55:15,070 --> 00:55:20,080 Because if we don't look at the population and Thomas Moses and David Ricardo and others were 468 00:55:20,080 --> 00:55:26,080 very influential in sort of pushing that thought and may be partially right about the history, 469 00:55:26,080 --> 00:55:30,100 but they were clearly not right about the time they were living through. 470 00:55:30,100 --> 00:55:39,550 And that led essentially to the abolishment of the poor laws, which was the only sort of source of relief for people who lost their jobs at the time. 471 00:55:39,550 --> 00:55:47,320 And that sort of just exacerbated the social unrest, and we actually know better this time around them, 472 00:55:47,320 --> 00:55:59,260 which is that it's great for those of you that today's the first day at Oxford to join us for drinks afterwards and celebrate your arrival. 473 00:55:59,260 --> 00:56:03,790 For those of you that have been haemorrhaging, you know, we don't have drinks after every event, 474 00:56:03,790 --> 00:56:10,060 so this is a particular treat today, and it's clearly been a treat to hear Karl. 475 00:56:10,060 --> 00:56:17,560 I'm absolutely delighted after having started the group seven years ago that it's reached the stage of 476 00:56:17,560 --> 00:56:23,980 global resonance and that calls produced not only a whole series of of great papers with his colleagues, 477 00:56:23,980 --> 00:56:27,400 but also this book published by Princeton University Press. 478 00:56:27,400 --> 00:56:34,670 This is also the only time you can be able to buy a book that costs twenty four pounds, something for fifteen pounds brand new. 479 00:56:34,670 --> 00:56:56,862 So I encourage you to do that, to socialise and to meet with Carl, and thanks to you all for coming.