WEBVTT - LIVE: We Are Not Machines - with Sarah O'Connor 0:00:15.410 --> 0:00:22.370 Pushkin. Tim here with an exciting announcement about the Cautionary Club, 0:00:22.690 --> 0:00:26.850 our Patreon community for Cautionary Tales listeners. So many of 0:00:26.850 --> 0:00:30.210 you joined our live table read earlier this year that 0:00:30.330 --> 0:00:33.610 we're doing it again. Join me and I production team 0:00:33.770 --> 0:00:37.690 for a live reading of an unreleased episode about the 0:00:37.730 --> 0:00:43.410 people who almost invented the iPhone fifteen years early and 0:00:43.450 --> 0:00:46.770 the surprising reasons they failed. This will be a chance 0:00:46.850 --> 0:00:49.010 for you to see how the stories we tell are 0:00:49.010 --> 0:00:53.490 developed in real time and ask your burning questions about 0:00:53.490 --> 0:00:57.010 Cautionary Tales. It's on the eighth of July at five 0:00:57.050 --> 0:01:01.090 pm UK or noon Eastern. If you join the Patreon 0:01:01.170 --> 0:01:03.770 before then, or if you're already a member, you'll get 0:01:03.770 --> 0:01:09.810 your exclusive invitation. Head to Patreon dot com slash Cautionary 0:01:09.850 --> 0:01:13.450 Club that Patreon p A t R e O N 0:01:13.770 --> 0:01:25.490 dot com slash Cautionary Club. Missus Goodair is enjoying a 0:01:25.570 --> 0:01:29.690 leisurely breakfast while she reads the morning paper. It's April 0:01:29.930 --> 0:01:34.410 eighteen twelve, and the headlines are dominated by Wellington's military 0:01:34.410 --> 0:01:40.010 successes in Spain. A strange noise jolts her from her reading. 0:01:40.770 --> 0:01:44.650 It sounds like the iron gates to her home rattling. 0:01:46.650 --> 0:01:49.730 She rushes upstairs to get a better view and sees 0:01:49.770 --> 0:01:53.410 the mob for it is a mob. More clearly, they're 0:01:53.410 --> 0:01:58.050 throwing stones, making lude gestures, and shouting for her to 0:01:58.130 --> 0:02:02.250 open the gates. Two men appear to be wearing dresses, 0:02:02.690 --> 0:02:08.450 proclaiming loudly to be General Lud's wives. She knows exactly 0:02:08.450 --> 0:02:12.050 what this is about. It's about what's in her husband's 0:02:12.130 --> 0:02:16.570 mill next door. So why are they at her house? 0:02:18.010 --> 0:02:22.450 Her husband, John Goodair, is out of town. She hovers 0:02:22.450 --> 0:02:26.770 at the window, unsure of what to do. The mob 0:02:26.850 --> 0:02:30.450 leaders grow impatient. They shout to their comrades that it's 0:02:30.490 --> 0:02:33.770 time to move on and begin marching through the streets 0:02:33.810 --> 0:02:37.970 of Stockport. As they weave through the town, they smash 0:02:38.050 --> 0:02:42.890 windows and break down factory doors. By noon, the mob 0:02:42.930 --> 0:02:48.970 has swelled over two thousand men. Soon they run out 0:02:48.970 --> 0:02:52.530 of targets, so they turned back to where they began, 0:02:53.490 --> 0:02:58.250 John Goodair's enormous mill, the place many of them have 0:02:58.330 --> 0:03:02.410 spent their working lives. Once a symbol of pride in 0:03:02.450 --> 0:03:06.650 their craft, now a symbol of everything they feel is 0:03:06.770 --> 0:03:11.850 wrong with the industrialized world. They break down the doors, 0:03:12.490 --> 0:03:16.970 smash through the windows, and destroy all they can inside 0:03:17.690 --> 0:03:22.810 every last frame, every last spindle. The mob has grown 0:03:23.130 --> 0:03:26.250 so large that even the iron gates to the Good 0:03:26.250 --> 0:03:31.970 Air Home can't hold them back. Missus Goodair has already 0:03:31.970 --> 0:03:38.050 gone a wise move. The Luddites burn the house to 0:03:38.130 --> 0:03:43.810 the ground. I'm Tim Harford and you're listening to Cautionary 0:03:43.850 --> 0:04:08.770 Tales live at the Bristol Festival of Economics. Loyal listeners 0:04:08.770 --> 0:04:11.690 will remember this is not the first time the Luddites 0:04:11.770 --> 0:04:15.090 have featured on this podcast. Two years ago I told 0:04:15.130 --> 0:04:18.730 you about the failure of the Luddite Revolution, the weavers 0:04:18.770 --> 0:04:24.010 who were incensed that automation was devaluing their craft in general, 0:04:24.090 --> 0:04:29.570 Lud's rage against the machines. Today, the Ludites seem more 0:04:29.610 --> 0:04:34.730 relevant than ever as the so called Ai revolution threatens 0:04:34.770 --> 0:04:38.890 to overturn the world of work. So should we be 0:04:39.130 --> 0:04:43.850 smashing up large language models or is resisting as futile 0:04:43.850 --> 0:04:46.410 as it was for the Luddites? Who's going to be 0:04:46.490 --> 0:04:51.730 most affected? And have economists been asking the wrong questions 0:04:51.890 --> 0:04:57.050 all along? I'm joined by a very special guest, Sarah O'Connor. 0:04:57.250 --> 0:05:00.810 She's a Financial Times journalist with a specialism in technology 0:05:00.890 --> 0:05:04.730 and work, and she's written a book called we are 0:05:04.770 --> 0:05:09.490 not machines the fight for the future of work. She 0:05:09.650 --> 0:05:12.610 is the perfect person to guide us through what's coming next, 0:05:13.050 --> 0:05:15.690 and she's got plenty of cautionary tales to tell us. 0:05:16.450 --> 0:05:18.890 Welcome Sarah, Hello Tim, Thank you for having me. 0:05:19.010 --> 0:05:22.010 Oh it's a pleasure. We've got to get something out 0:05:22.010 --> 0:05:25.770 of the way first. So, Sarah O'Connor, have you seen 0:05:25.770 --> 0:05:31.570 the Terminator movies yet? No, you're gonna have to watch them. 0:05:31.810 --> 0:05:33.490 It's a point of principle. 0:05:33.090 --> 0:05:34.810 That, yeah, it now feels like stubbornness. 0:05:34.970 --> 0:05:37.250 Yes, so you might have noticed that my name bears 0:05:37.250 --> 0:05:40.730 a faint resemblance to the heroine of the Terminator films. 0:05:40.810 --> 0:05:43.290 Growing up, occasionally people made Terminator jokes to me, and 0:05:43.290 --> 0:05:45.810 as a result, I never watched them on principle because 0:05:45.850 --> 0:05:48.290 it annoyed me. And so about a decade ago, I 0:05:48.370 --> 0:05:53.210 went accidentally sort of globally viral when I very simply 0:05:53.290 --> 0:05:56.370 tweeted a news story that i'd seen come out, which 0:05:56.450 --> 0:05:58.810 was that a robot had killed a worker in a 0:05:58.850 --> 0:06:02.730 Volkswagen factory, and didn't realize that this would cause the 0:06:02.730 --> 0:06:06.650 Internet to break. So yeah, so people just found it 0:06:06.770 --> 0:06:10.810 very hilarious that someone called Sarah O'Connor had written about robots, 0:06:10.850 --> 0:06:13.970 and for about a decade people have occasionally sent me, 0:06:14.570 --> 0:06:18.330 you know, terminator gifts posters of the Terminator films with 0:06:18.370 --> 0:06:21.730 my face photoshopped in, and as a result, no, I 0:06:21.730 --> 0:06:22.650 have not watched those. 0:06:22.810 --> 0:06:27.850 Okay, So the book is it's fantastic, we are not machines. 0:06:28.330 --> 0:06:30.690 What's the headline? What are you basically driving out with 0:06:30.730 --> 0:06:31.210 this book? 0:06:31.410 --> 0:06:34.130 I think what I'm trying to say in the book 0:06:34.210 --> 0:06:37.810 is that, you know, we're all sort of drowning in 0:06:37.810 --> 0:06:43.050 information predictions about AI and what it might do to 0:06:43.090 --> 0:06:44.810 the world of work, and a lot of those are 0:06:44.850 --> 0:06:49.010 coming from either economists or the big tech executives who 0:06:49.090 --> 0:06:54.370 have created large language models and are marketing them. 0:06:53.490 --> 0:06:55.970 And both these sources are unreliable for different reasons. 0:06:56.450 --> 0:06:59.490 Yeah, the tech guys have something to sell, right, and 0:06:59.530 --> 0:07:04.970 so they have their own motivations. And then you know, economists, 0:07:05.050 --> 0:07:07.170 I think are trying their best to figure out what's 0:07:07.170 --> 0:07:10.250 going on. But the pay which they're doing that I 0:07:10.250 --> 0:07:12.850 think is very feels very abstract to me as someone 0:07:12.930 --> 0:07:15.170 you know, I'm a reporter at heart, So my favorite 0:07:15.210 --> 0:07:17.250 thing to do is to like get my notebook, put 0:07:17.250 --> 0:07:19.970 my boots on and like go and meet people and 0:07:20.490 --> 0:07:22.810 stand outside the factory gates and talk to people about 0:07:22.810 --> 0:07:25.730 what's actually happening. And so over the past few years, 0:07:25.730 --> 0:07:28.250 I've just felt very frustrated that everything that I'm hearing 0:07:28.290 --> 0:07:30.970 about this is coming from people who are either looking 0:07:30.970 --> 0:07:34.970 at spreadsheets and models or who have a product to sell, 0:07:35.370 --> 0:07:37.050 and I just wanted to hear less about what these 0:07:37.130 --> 0:07:39.210 men say is going to happen, and more about what's 0:07:39.290 --> 0:07:42.410 already actually happening from the people that it's actually happening too. 0:07:42.810 --> 0:07:45.370 So in the book, I basically went to find people 0:07:45.610 --> 0:07:48.570 and places and workplaces that are on the front lines 0:07:48.610 --> 0:07:52.690 of what's already happening. So people who are encountering autonomous 0:07:52.770 --> 0:07:56.370 robots in their workplaces, or are working with or for 0:07:57.090 --> 0:08:00.170 or around large luggage models, and they're starting to change 0:08:00.410 --> 0:08:02.530 the way they work. And when I did that, what 0:08:02.610 --> 0:08:05.370 I realized was that you, no, I wasn't really hearing 0:08:05.530 --> 0:08:09.330 this very utopian story that this is liberating me, this 0:08:09.570 --> 0:08:13.450 is freeing me from dull, dirty, dangerous work and making 0:08:13.810 --> 0:08:16.290 everything fantastic. But nor was I really hearing a lot 0:08:16.330 --> 0:08:19.010 of stories that were like my job has been completely 0:08:19.090 --> 0:08:21.850 wiped out. What I was hearing instead were stories about 0:08:22.410 --> 0:08:25.890 quite profound changes to the nature of people's jobs and 0:08:25.930 --> 0:08:28.690 how they felt about their jobs and whether they enjoyed 0:08:28.730 --> 0:08:30.770 their jobs or not. And some of those were quite 0:08:31.170 --> 0:08:32.250 cautionary tales. 0:08:32.450 --> 0:08:32.690 You know. 0:08:32.890 --> 0:08:35.730 There were a lot of people who were saying, actually, 0:08:35.810 --> 0:08:38.370 what's happening to me is almost the opposite of this 0:08:38.490 --> 0:08:41.570 idea that technology is going to give us space to 0:08:41.610 --> 0:08:44.170 be more human. People were finding themselves sort of crunched 0:08:44.210 --> 0:08:47.890 into systems that were paced by machines run by machines, 0:08:47.930 --> 0:08:52.530 in which judgment was being overtaken by machines, in which 0:08:52.570 --> 0:08:55.970 creativity was being sort of compressed or contained by machines. 0:08:56.250 --> 0:08:58.370 And I thought, actually, if that's what's beginning to happen 0:08:58.410 --> 0:09:00.570 to people, then everyone should know about that, and we 0:09:00.570 --> 0:09:03.450 should try and talk about why that's happening and whether 0:09:03.490 --> 0:09:06.170 there is a way to avoid that, because fundamentally, I 0:09:06.210 --> 0:09:07.490 don't think that's what anybody wants. 0:09:07.650 --> 0:09:09.490 Before we sort of take a trip to to the 0:09:09.490 --> 0:09:12.370 cutting edge, I just I wanted to go back a 0:09:12.410 --> 0:09:15.090 couple of hundred years we began with the ludd heightes. 0:09:15.170 --> 0:09:17.810 Are they still relevant? Do people misunderstand the story of 0:09:17.850 --> 0:09:18.410 the loudtes? 0:09:18.530 --> 0:09:18.690 Yeah? 0:09:18.690 --> 0:09:20.650 I think they are still relevant. I mean, Luddite is 0:09:20.770 --> 0:09:23.490 just now used as a sort of a flippant term 0:09:23.530 --> 0:09:26.730 for somebody who's sort of reflexively anti technology and just 0:09:26.730 --> 0:09:29.290 doesn't like any new tech. But actually the ludd Heites, 0:09:29.410 --> 0:09:32.850 they weren't really fighting against technology full stop. They were 0:09:33.010 --> 0:09:36.810 fighting against the way in which machines were being put 0:09:36.850 --> 0:09:39.810 to a new use, which was to cut them out 0:09:39.810 --> 0:09:43.610 of the equation. These sort of knitting and wide frame machines. 0:09:43.650 --> 0:09:47.010 They weren't displacing work entirely. What they did was they 0:09:47.050 --> 0:09:50.490 allowed people who had no skill and no experience to 0:09:50.570 --> 0:09:54.090 make a less good quality product. And in that sense, 0:09:54.130 --> 0:09:56.290 as we might come onto it, actually has an awful 0:09:56.330 --> 0:09:59.690 lot of similarities with some of the technologies that we're 0:09:59.730 --> 0:10:00.330 seeing today. 0:10:00.490 --> 0:10:03.290 Instead of knitting a stocking, you knit a massive bit 0:10:03.330 --> 0:10:05.930 of cloth that you can make up punch stockings out of, 0:10:05.970 --> 0:10:07.730 and you just cut them into strips and then you 0:10:07.730 --> 0:10:09.850 make the stockings. The stockings the terrible they fall apart. 0:10:10.090 --> 0:10:13.890 So you've suddenly got interesting, well paid jobs being replaced 0:10:13.890 --> 0:10:19.010 by boring, badly paid jobs. You've got expensive, high quality 0:10:19.610 --> 0:10:22.130 work being replaced by cheap, crappy work. 0:10:22.330 --> 0:10:25.450 Even if they were lower quality. They were also a 0:10:25.530 --> 0:10:27.250 much lower price, and that I think that was a 0:10:27.250 --> 0:10:29.210 deal that a lot of consumers were willing to make. 0:10:29.690 --> 0:10:33.330 So let's talk about warehouses. The picture I have in 0:10:33.370 --> 0:10:35.290 my mind, or the picture I had in my mind 0:10:35.370 --> 0:10:38.290 before I read the book, was, in fact, I realized 0:10:38.330 --> 0:10:42.210 a picture painted for me by you about ten years ago. 0:10:42.690 --> 0:10:45.770 One of the things that some workers in these had 0:10:45.930 --> 0:10:49.570 was they had this earpiece, the Jennifer unit, basically just 0:10:49.610 --> 0:10:51.650 a voice in their ear telling them where to go 0:10:51.690 --> 0:10:54.450 in the warehouse, what to pull off the shelves. You 0:10:54.490 --> 0:10:56.890 don't need them to think about the best way to 0:10:56.930 --> 0:10:59.770 get around the warehouse or to remember anything, just tobay 0:10:59.770 --> 0:11:01.930 the voice in your head. So this sort of creation 0:11:02.010 --> 0:11:05.610 of these flesh robots. So I was reading this article 0:11:05.650 --> 0:11:07.570 by you about ten years ago and thinking that that 0:11:07.690 --> 0:11:10.770 sounds bad. But of course that's ten years ago. So 0:11:11.250 --> 0:11:15.890 how accurate is that image of what is now happening? 0:11:15.890 --> 0:11:16.570 What's changed? 0:11:17.090 --> 0:11:19.290 Yeah, so quite a lot has changed. And actually I 0:11:19.330 --> 0:11:22.770 think that article from ten years ago and that sort 0:11:22.810 --> 0:11:24.850 of reporting that I've done, And just to clarify that 0:11:24.970 --> 0:11:27.450 Jennifer unit isn't something that's used in Amazon warehouses, but 0:11:27.490 --> 0:11:30.450 it is used in some other warehouses really informed the 0:11:30.490 --> 0:11:32.810 way I used to think about technology. So, yeah, I 0:11:32.850 --> 0:11:34.450 had met and interviewed a lot of people over the 0:11:34.490 --> 0:11:36.330 years who worked in jobs in which they were basically 0:11:36.410 --> 0:11:40.850 expected to work like robots. I used to think, well, 0:11:41.090 --> 0:11:43.330 you know, bring on the down robots. Then you know, like, 0:11:43.370 --> 0:11:45.650 let's get real robots in to do these jobs, like 0:11:45.690 --> 0:11:48.530 this is a waste of kind of human potential, and 0:11:48.890 --> 0:11:51.490 this is a great example of where automation would be brilliant. 0:11:51.530 --> 0:11:53.410 And so I used to be quite a techno optimist 0:11:53.450 --> 0:11:55.570 in the sense that I thought there were huge numbers 0:11:55.570 --> 0:11:58.570 of roles that actually could and should be automated. So 0:11:58.610 --> 0:12:02.890 what's happened now in Amazon warehouses, at least in the 0:12:02.890 --> 0:12:06.050 one that I went to visit, is that the robots 0:12:06.090 --> 0:12:09.170 have arrived, but not really in the way that I had. 0:12:09.410 --> 0:12:11.970 So I'd basically imagined like a one for one swap, right, 0:12:12.010 --> 0:12:14.450 so instead of a person walking around, you'd have like 0:12:14.450 --> 0:12:17.290 a kind of human humanoid robot doing it. 0:12:17.730 --> 0:12:20.170 But it's never actually like that, is it. It's never 0:12:20.250 --> 0:12:24.250 ever like that, Like, a robot accountant is basically Microsoft exell. 0:12:24.290 --> 0:12:27.010 A robot accountant is not C three post sitting in 0:12:27.010 --> 0:12:29.050 your chair top. 0:12:28.970 --> 0:12:31.570 No, exactly, and it's not how it works in Amazon either. 0:12:31.650 --> 0:12:35.050 So I went to visit a new Amazon warehouse which 0:12:35.370 --> 0:12:38.490 has their sort of latest automation technology. And the way 0:12:38.530 --> 0:12:41.010 it works now is that rather than you, as a 0:12:41.050 --> 0:12:43.610 worker walking around for like maybe ten miles a day 0:12:43.730 --> 0:12:46.970 just weaving between all of these shelves and picking things off, 0:12:47.490 --> 0:12:50.250 now on each floor there's a big fence and around 0:12:50.290 --> 0:12:54.570 the perimeter of that fence, the workers stand stationary in 0:12:54.610 --> 0:12:56.530 one place and inside the fence of the robots. But 0:12:56.530 --> 0:12:58.930 the robots are basically like giant rumbers, and what they 0:12:58.970 --> 0:13:01.210 do is they drive around inside this perim defence and 0:13:01.210 --> 0:13:04.130 they pick up the shelves and they bring the shelves 0:13:04.170 --> 0:13:06.810 to the workers. And so you now, as a worker, 0:13:06.850 --> 0:13:08.770 you stand in one place and a robot brings you 0:13:08.810 --> 0:13:12.250 a shirt, and then a light illuminates which part of 0:13:12.250 --> 0:13:14.250 the shelf you need to look at, and a screen 0:13:14.330 --> 0:13:15.810 gives you a picture of what the thing is you 0:13:15.850 --> 0:13:17.770 need to pick off. And I tried this out. I 0:13:17.770 --> 0:13:19.610 think I had to pick off like a mobile phone 0:13:19.650 --> 0:13:21.290 case with a dolphin on it, so you know, you 0:13:21.330 --> 0:13:24.130 pick it out, and then another light illuminates which box 0:13:24.170 --> 0:13:25.570 that you need to put it into and you push 0:13:25.570 --> 0:13:28.370 a big button and then that robot takes that shelf away, 0:13:28.410 --> 0:13:30.250 and the next robot's already queuing up for you with 0:13:30.290 --> 0:13:31.930 the next shelf, and you do the same thing. 0:13:32.130 --> 0:13:33.370 How long did you do it? For? 0:13:33.650 --> 0:13:34.290 Two minutes? 0:13:34.530 --> 0:13:36.490 Okay? How long? How long do they do it for? 0:13:36.610 --> 0:13:39.570 Ten hours a day, forty hours a week with two 0:13:39.650 --> 0:13:42.170 thirty minute breaks? And so this is not what I 0:13:42.250 --> 0:13:45.370 had imagined when I sort of cheered on the robots 0:13:45.370 --> 0:13:47.530 and thought this would be great. You know, in some 0:13:47.570 --> 0:13:50.050 ways I think that job is better. It is less 0:13:50.090 --> 0:13:52.930 physically demanding than walking around for ten hours a day. 0:13:53.330 --> 0:13:55.930 Amazon says that it's safer so they have, you know, 0:13:56.090 --> 0:13:59.250 fewer injuries and accidents, and I think that's probably true. 0:13:59.770 --> 0:14:02.210 But if anything, I think it's also become more boring 0:14:02.530 --> 0:14:05.970 and more monotonous. And also it's still not that easy 0:14:06.530 --> 0:14:08.410 on the human body to stand up for ten hours 0:14:08.450 --> 0:14:11.810 a day, and because they're now much more productive, right 0:14:11.850 --> 0:14:14.770 because you're not walking around from one place to the next, 0:14:14.770 --> 0:14:17.370 and so the products are coming at you, so you're bending, 0:14:17.410 --> 0:14:20.250 you're lifting, you're twisting over and over and over again. 0:14:20.290 --> 0:14:22.410 So I think it's still quite quite a physically demanding 0:14:22.530 --> 0:14:25.690 job and arguably like a bit more monotonous and a 0:14:25.690 --> 0:14:26.530 bit lonelier. 0:14:27.010 --> 0:14:28.850 Presumably there are people who did the old job and 0:14:28.890 --> 0:14:30.450 now do the new job, So what do they think. 0:14:31.050 --> 0:14:34.970 Well, I interviewed one worker who had volunteer to transfer 0:14:35.050 --> 0:14:39.370 from a manual, old fashioned warehouse to a new one, 0:14:39.530 --> 0:14:41.530 and he wanted to go back. He said, actually, I 0:14:41.530 --> 0:14:44.090 don't like working with the robots. It's too intense, it's 0:14:44.130 --> 0:14:46.810 too exhausting, it's really lonely. I don't get to talk 0:14:46.850 --> 0:14:49.730 to anyone anymore. And he yeah, he's now transferred back 0:14:49.770 --> 0:14:51.450 to his old warehouse. 0:14:51.730 --> 0:14:56.050 By way of contrast, you also visited a mine in Sweden, 0:14:56.330 --> 0:14:59.090 So how is automation taking off there? 0:14:59.890 --> 0:15:02.730 Yeah, so this mine in Sweden was right up near 0:15:02.770 --> 0:15:06.290 the fringe of the Arctic Circle and their mine copper 0:15:06.330 --> 0:15:10.050 there and various other minerals, and it is one of 0:15:10.090 --> 0:15:12.970 the most sort of technologically advanced mines in the world, 0:15:13.010 --> 0:15:14.930 or so they say. So it used to be that 0:15:14.970 --> 0:15:19.610 miners would drive vehicles around below. Grounds are very deep down. 0:15:19.930 --> 0:15:22.650 It's horrible down there. It's like really dark, it's really 0:15:22.770 --> 0:15:26.730 humid and claustrophobic. And now the vehicles are autonomous, so 0:15:26.770 --> 0:15:30.530 they drive themselves, and the miners sit in a control room. 0:15:30.850 --> 0:15:32.850 They can take over the vehicles if they need to, 0:15:32.930 --> 0:15:34.970 but a lot of the time they're just they're sitting there, 0:15:35.010 --> 0:15:37.250 They're watching what's happening on big screens. They're in a 0:15:37.250 --> 0:15:40.850 comfy chair, and they're listening to Spotify, they're listening to 0:15:41.490 --> 0:15:43.650 the ice hockey. And the ones that I spoke to 0:15:44.090 --> 0:15:47.250 were quite satisfied with this change in their working conditions. 0:15:47.490 --> 0:15:50.330 Is the fact that the Swedish mineworkers are happy and 0:15:50.370 --> 0:15:54.650 the Amazon workers are at best ambivalent about the way 0:15:54.690 --> 0:15:57.250 their job works. What explains the difference. 0:15:57.610 --> 0:15:59.930 So I think there's a few differences. One difference is that, 0:16:00.530 --> 0:16:02.690 you know, the reason that in Amazon they don't have 0:16:02.770 --> 0:16:05.570 humanoids walking around doing the whole thing is that that 0:16:05.730 --> 0:16:09.090 technology is not ready yet and so in some ways 0:16:09.210 --> 0:16:12.530 are still kind of plugged into a system which is 0:16:12.970 --> 0:16:17.250 like part robot, part human, but is increasingly being paced 0:16:17.330 --> 0:16:20.370 by the capabilities of the robots, Whereas I guess for 0:16:20.450 --> 0:16:23.690 the miners, they do not need to be in the 0:16:23.690 --> 0:16:26.890 cab anymore, and so that means that there's the potential 0:16:26.890 --> 0:16:30.410 for a much kind of greater improvement in their health 0:16:30.410 --> 0:16:32.210 and safety. But as well as that, I think there 0:16:32.290 --> 0:16:35.370 is a governance thing. So I don't know how much 0:16:35.410 --> 0:16:38.770 everyone knows about Sweden, but they have a very distinctive 0:16:38.890 --> 0:16:42.690 kind of labor market whereby trade unions are very powerful, 0:16:43.330 --> 0:16:46.250 and there is a rule that any new technology has 0:16:46.290 --> 0:16:49.290 to be bargained collectively about before it happens. And so 0:16:50.090 --> 0:16:52.330 before they introduced this new technology, they had to sit 0:16:52.370 --> 0:16:54.210 down with the workers and they had to talk about 0:16:54.210 --> 0:16:56.570 it and how it would go and what the workers' 0:16:56.610 --> 0:16:59.890 concerns were and what the company wanted to get out 0:16:59.890 --> 0:17:03.330 of it. And you know, more broadly in Scandinavian countries, 0:17:03.370 --> 0:17:06.250 what you find is that when people answer surveys about 0:17:06.250 --> 0:17:09.210 how they feel about technology, people are like much more 0:17:09.210 --> 0:17:11.690 positive than they are in the UK or in the US, 0:17:11.810 --> 0:17:13.330 and I think that's because they have a sense that 0:17:13.370 --> 0:17:15.410 they will have a seat at the table when those 0:17:15.410 --> 0:17:16.330 decisions get made. 0:17:16.610 --> 0:17:20.450 Thank you, Sarah. You're listening to a special cautionary conversation 0:17:20.610 --> 0:17:24.370 recorded live at the Bristol Festival of Economics. And after 0:17:24.410 --> 0:17:28.530 the break, my guests Sarah O'Connor will be telling a 0:17:28.690 --> 0:17:42.770 cautionary tale about language translators at artificial intelligence. Stay with us, 0:17:45.570 --> 0:17:49.050 we're back this is the Bristol Festival of Economics. I 0:17:49.090 --> 0:17:53.010 am Tim Harford and my special guest is Financial Times 0:17:53.090 --> 0:17:59.930 journalist and author Sarah O'Connor. So, Sarah, perhaps the canaries 0:17:59.970 --> 0:18:03.850 in the coal mine are the translators. When I thought 0:18:03.890 --> 0:18:07.210 about translators, I just thought, oh, you guys are just 0:18:07.250 --> 0:18:09.490 all You've lost all your jobs because of Google Translate. 0:18:09.570 --> 0:18:13.290 It's all gone. But actually your the story you tell 0:18:13.290 --> 0:18:15.010 in the book is is a lot more interesting and 0:18:15.050 --> 0:18:15.890 more complex than that. 0:18:16.170 --> 0:18:20.210 Yeah. So I interviewed some translators for my book. One 0:18:20.210 --> 0:18:22.490 of them is a guy called Peter. He lives in 0:18:22.490 --> 0:18:26.370 the Czech Republic and he translates subtitles for TV shows, 0:18:26.930 --> 0:18:28.770 which is a great job. I mean, he told me 0:18:28.810 --> 0:18:31.210 lots of things about why it's more difficult than you 0:18:31.290 --> 0:18:34.450 might imagine to do that. So, for example, you know, 0:18:34.490 --> 0:18:36.530 in English, we just have one way of saying you, 0:18:36.730 --> 0:18:39.610 but in lots of languages, including Zech, there's like a 0:18:39.690 --> 0:18:42.010 formal tense of you and an informal tense of you. 0:18:42.130 --> 0:18:45.570 And so when you're writing the translated subtitles, you have 0:18:45.650 --> 0:18:48.850 to think like, how well do these characters know each other? 0:18:48.890 --> 0:18:51.010 Like what tense would they be speaking in? And that 0:18:51.090 --> 0:18:53.170 might change, you know, through the course of an episode 0:18:53.250 --> 0:18:55.090 or through the course of a series, and so there 0:18:55.130 --> 0:18:57.250 are lots of kind of interesting judgment calls that you 0:18:57.730 --> 0:19:00.370 have to make. And then, you know, translating jokes is 0:19:00.410 --> 0:19:04.130 another thing that's incredibly difficult because so many jokes are 0:19:04.170 --> 0:19:07.090 like culture specific or they're like plays on language. 0:19:07.170 --> 0:19:09.650 We're in a bookshop where the guests of war systems 0:19:09.690 --> 0:19:12.690 in Bristol. There must be some asterisks and obeliqs somewhere 0:19:12.690 --> 0:19:13.290 in the bookshop. 0:19:13.370 --> 0:19:14.730 Yeah, they're the perfect example. 0:19:14.810 --> 0:19:15.010 Right. 0:19:15.090 --> 0:19:20.090 So Obelix's dog in the original French is called ede feaks, 0:19:20.210 --> 0:19:24.290 which means obsession, and the brilliant English translation is dogmatics. 0:19:24.850 --> 0:19:27.210 You know. So like, there's so much sort of creativity 0:19:27.250 --> 0:19:30.050 and humor involved in being a really good translator, and 0:19:30.090 --> 0:19:31.930 so like, this is a great job for people that 0:19:32.130 --> 0:19:34.490 enjoy doing that sort of thing. But what's happened to 0:19:34.570 --> 0:19:37.010 him and to people like him, is that he has 0:19:37.050 --> 0:19:40.410 not lost his job entirely. Like, it's actually quite hard 0:19:40.450 --> 0:19:43.290 to get a machine to translate things with that level 0:19:43.330 --> 0:19:47.650 of cultural knowledge and understanding. So, just to explain slightly 0:19:47.690 --> 0:19:50.130 the way a lot of translation works is a lot 0:19:50.130 --> 0:19:53.610 of translators of freelancers, and there are agencies that will 0:19:53.610 --> 0:19:57.850 take work from you know, a TV studio and then 0:19:57.890 --> 0:20:00.290 parcel it out to freelance translators. And what those agencies 0:20:00.330 --> 0:20:03.930 have started to do is to take something, get it 0:20:03.970 --> 0:20:06.890 translated by a large language model or a different kind 0:20:06.930 --> 0:20:10.050 of machine translation service, and then it to the translator 0:20:10.090 --> 0:20:11.970 and say, hey, can you just check that this is 0:20:12.050 --> 0:20:14.210 right and maybe like just finesse it a little bit, 0:20:14.250 --> 0:20:16.890 polish it, make it sound a bit more human. And 0:20:16.930 --> 0:20:18.930 this has become the kind of bete noir of lots 0:20:18.970 --> 0:20:23.090 of translators, because what they say is that it's a 0:20:23.650 --> 0:20:26.290 you know, you're expected to do this much faster and 0:20:26.330 --> 0:20:28.530 be you're expected to do it for much less money. 0:20:29.010 --> 0:20:31.210 But in fact, if you care about quality, it's very 0:20:31.210 --> 0:20:35.330 difficult to actually check a translation is accurate, and then 0:20:35.370 --> 0:20:38.010 even things like trying to make it sound a bit 0:20:38.050 --> 0:20:40.330 better like that's actually what the translators told me was, 0:20:40.370 --> 0:20:42.930 it's actually quite hard to do when there's already like 0:20:42.930 --> 0:20:45.450 an answer in front of you, and it feels much 0:20:45.530 --> 0:20:49.050 less creative and more kind of cumbersome and less enjoyable. 0:20:49.170 --> 0:20:52.210 And so a job that that was sort of creative 0:20:52.250 --> 0:20:58.730 and challenging and interesting, has become faster, harder, more monotonous, 0:20:58.770 --> 0:21:03.330 feels more mechanical, and fundamentally kind of less satisfying. And 0:21:03.370 --> 0:21:05.050 I think that is a real cautionary tale. And when 0:21:05.050 --> 0:21:06.650 I was speaking to them, it made me think a 0:21:06.690 --> 0:21:09.890 lot of the luod ites, because what the translators will 0:21:09.890 --> 0:21:12.530 tell you is that the end product is less good 0:21:12.610 --> 0:21:15.530 quality than There are even some studies that have checked this. 0:21:15.730 --> 0:21:18.130 So there's a big study that looked at a bunch 0:21:18.170 --> 0:21:22.210 of subtitles from TV shows before and after the introduction 0:21:22.330 --> 0:21:26.090 of machine translation post editing, and the quality has kind 0:21:26.090 --> 0:21:27.010 of deteriorated. 0:21:27.370 --> 0:21:28.370 It's the stockings again. 0:21:28.490 --> 0:21:29.930 It's the stockings again, exactly. 0:21:30.210 --> 0:21:33.170 And I guess partly obviously we care about the quality 0:21:33.170 --> 0:21:36.930 of jobs. That's important. But I guess one of the 0:21:36.970 --> 0:21:40.850 interesting questions from the point of view of the consumer, 0:21:41.370 --> 0:21:45.210 because the whole idea here is even if individuals get 0:21:45.250 --> 0:21:50.890 worse jobs, the consumer gets more choice, more quality, lower prices. 0:21:51.810 --> 0:21:56.130 I always do sort of wander is the product actually 0:21:56.170 --> 0:22:00.930 better and will the market deliver kind of the right 0:22:01.010 --> 0:22:02.970 trade offs or are we actually just going to make 0:22:03.250 --> 0:22:06.130