Latest / Trumponomics / The Key to Making AI a Benefit, Not a Hazard

Transcript

The Key to Making AI a Benefit, Not a Hazard · Trumponomics

0:10
Speaker 1: Hello, and welcome to Stefanowie's the podcast that brings you
0:13
Speaker 1: the global economy, and we're dedicating this week's episode to
0:16
Speaker 1: a powerful book on a subject most of us have
0:19
Speaker 1: been thinking a lot about recently, how new technology can
0:23
Speaker 1: change the world. Chat GPT came out the end of
0:26
Speaker 1: last year, had a million users in the space of
0:29
Speaker 1: five days, and that was pretty cool. Then in March,
0:33
Speaker 1: chat GPT four came out. Now that was a whole
0:36
Speaker 1: lot better, but also alarming. Journalists, lawyers, accountants, teachers could
0:42
Speaker 1: all see how it could not only help them do
0:45
Speaker 1: their jobs but make them redundant. Not sometime in the future,
0:49
Speaker 1: but next month. Should we worry then about where this
0:53
Speaker 1: will lead? Well, the standard version of economic history says
0:57
Speaker 1: not really. It tells a story where I and again
1:00
Speaker 1: people fear technology will make the world worse, think of
1:04
Speaker 1: those Luddites smashing up machines. But in the end it's better. Yeah,
1:09
Speaker 1: there are adjustment costs, a bunch of people lose their jobs, maybe,
1:13
Speaker 1: but overall, the majority of people get new opportunities, more
1:17
Speaker 1: rewarding jobs, better lives. As the version of history I
1:20
Speaker 1: was taught in graduate school and have heard from fellow
1:23
Speaker 1: economists many many times since, and we've heard it again
1:26
Speaker 1: often in response to AI. To be scared of this
1:30
Speaker 1: new technology, to believe it will hurt workers, they say,
1:35
Speaker 1: you have to believe this technological revolution will be different
1:38
Speaker 1: from all those that have gone before. Now you might
1:41
Speaker 1: think that's true. This time is different. Chat GPT feels different,
1:46
Speaker 1: but it sets a high bar for being scared. After all,
1:49
Speaker 1: the horseless carriage also felt pretty different at the time.
1:53
Speaker 1: But two very distinguished economists have taken a fresh look
1:56
Speaker 1: at that history and decided the basic reassuring story about
2:01
Speaker 1: the past impact of technology on jobs and the quality
2:04
Speaker 1: of life of working people is wrong or seriously incomplete.
2:10
Speaker 1: So if their argument is right, we should be worried
2:13
Speaker 1: about the way AI will transform our economy and society
2:17
Speaker 1: because we are not now in a position to get
2:19
Speaker 1: the best out of it. Quite the opposite. That book
2:22
Speaker 1: is Power and Progress, Our thousand years Struggle over Technology
2:26
Speaker 1: and Prosperity, and the authors are darn Smoglu and Simon Johnson,
2:31
Speaker 1: both professors at the Massachusetts Institute of Technology. Simon's a
2:35
Speaker 1: former chief economist at the International Monetary Fund, and Darren
2:39
Speaker 1: won the John Bates Prize for the best US economist
2:42
Speaker 1: under forty a few years ago. He also wrote possibly
2:45
Speaker 1: the most widely read book of economic history in recent times,
2:49
Speaker 1: Why Nations Fail. And he's here with me now, Darren,
2:52
Speaker 1: Thank you very much. I'm really pleased we can have
2:54
Speaker 1: this conversation on Stephanomics.
2:56
Speaker 2: I am so happy to be here, and you give
2:58
Speaker 2: such a wonderful introduction. I don't think I have much
3:00
Speaker 2: to add, No, no.
3:01
Speaker 1: I think we have plenty to discuss, and I'm sorry
3:03
Speaker 1: about the long introduction, but I wanted to make clear
3:05
Speaker 1: why I thought the book was so important, and we
3:08
Speaker 1: do have I think a lot to unpack about the
3:11
Speaker 1: past and the present and what you hope might be
3:14
Speaker 1: a better future. But maybe we can start at that
3:17
Speaker 1: endpoint with what worries you right now about the impact
3:20
Speaker 1: that AI and similar technologies might have on the world.
3:25
Speaker 2: Well, I think you put it so well. We need
3:27
Speaker 2: to be concerned but not scared. I think this is
3:31
Speaker 2: a turning point. There are many transformative choices we have
3:37
Speaker 2: to make about the future of work, future of inequality,
3:39
Speaker 2: future of democracy. And the two worst positions we can
3:42
Speaker 2: take are to say everything's going to be fine, just
3:47
Speaker 2: let experts worry about it or be scared about killer robots.
3:52
Speaker 2: Both of them pacify us and push us not to
3:57
Speaker 2: take up our responsibility of trying to shape the technology,
4:01
Speaker 2: trying to get involved about decisions about the future of
4:04
Speaker 2: this technology.
4:05
Speaker 1: And I guess there is a lot there, but one
4:07
Speaker 1: piece of it that I was struck by because I
4:09
Speaker 1: feel like that was definitely many many of my economic
4:12
Speaker 1: history lessons had this sort of underlying premise. And you
4:15
Speaker 1: talk about the productivity bandwagon, so maybe you should explain that.
4:18
Speaker 2: Let me actually take a step back, Simon, and I
4:22
Speaker 2: are definitely not arguing that we haven't tremendously benefited from
4:26
Speaker 2: industrial technology and the scientific advances. Today we are so
4:30
Speaker 2: much more comfortable, so much more prosperous, so much healthier
4:33
Speaker 2: than people who lived three hundred years ago, and that's
4:35
Speaker 2: thanks to industrial technology and use of scientific knowledge in
4:39
Speaker 2: further improving technology in every aspect of our lives. What
4:43
Speaker 2: we are questioning that that was an automatic process. So
4:46
Speaker 2: the techno optimism that you so eloquently described at the
4:49
Speaker 2: beginning is that we don't need to take any drastic
4:53
Speaker 2: actions or become involved in shaping the future of technology
4:58
Speaker 2: because there is a very powerful in axorable, automatic process
5:01
Speaker 2: that's going to bring all sorts of good things to people.
5:03
Speaker 2: And at the center of it is what we call
5:05
Speaker 2: the productivity bandwagon. Because most of us earn our livings
5:09
Speaker 2: by supplying our labor. So the process via which technology
5:12
Speaker 2: is going to improve brings shared prosperity has to go
5:14
Speaker 2: through the labor market, which means wages have to improve.
5:17
Speaker 2: And the productivity bandwagon says that if productivity grows, if
5:23
Speaker 2: technological capabilities improve, that's going to create a very powerful
5:27
Speaker 2: force towards employers wanting more labor, and that raises wages
5:32
Speaker 2: and employment. If the productivity bandwagon breaks down or doesn't
5:37
Speaker 2: have many people on it, then shared prosperity would become
5:40
Speaker 2: a dream. And the real fear is that that's exactly
5:43
Speaker 2: what could happen with AI, and we've seen some of
5:45
Speaker 2: that happen with digital technologies over the last forty years anyway,
5:49
Speaker 2: and history says productivity bandwagon can work, but only if
5:54
Speaker 2: we create the right institutions and the right direction of technology.
5:57
Speaker 1: So what matters most for making an outcome better worse
6:00
Speaker 1: for workers?
6:01
Speaker 2: Perfect That's exactly the right question. And what Simon and
6:04
Speaker 2: I argue is that there are two pillars to it,
6:08
Speaker 2: and you can see them very clearly in most historical episodes. First,
6:12
Speaker 2: you need a direction of technology that doesn't just automate work.
6:15
Speaker 2: Automation is always going to be with us, but it
6:18
Speaker 2: doesn't just automate work, but at the same time creates
6:20
Speaker 2: new tasks, new capabilities, new things in which human labor
6:24
Speaker 2: can be productively used. And second, you need an institutional
6:27
Speaker 2: framework in which there are forces such as worker voice
6:31
Speaker 2: and worker power that induces employers to share some of
6:35
Speaker 2: the gains with workers. If either of those two things
6:38
Speaker 2: are broken down, then we're in trouble. If both of
6:42
Speaker 2: them break down, that's really damaging. And that's the age
6:45
Speaker 2: we are living in. There is no worker voice. AI
6:48
Speaker 2: is being used to sideline workers even more in the
6:50
Speaker 2: production process, and there isn't a democratic process that's actually
6:54
Speaker 2: contributing to it, to a sharing of prosperity, or to
6:59
Speaker 2: reshaping the direction of technology.
7:01
Speaker 1: You're talking about where we are now, which I think
7:02
Speaker 1: we should definitely get to. But for those who sort
7:05
Speaker 1: of feel like, oh, this sounds like people who are
7:08
Speaker 1: just seeing this new technology and fearing the worst, you know,
7:15
Speaker 1: I think at least we should show how it's rooted
7:17
Speaker 1: in that understanding of history that you mentioned, and maybe
7:21
Speaker 1: draw some contrasts when you're looking back, for example, at
7:23
Speaker 1: the nineteenth century. I mean, there's a lot of a
7:25
Speaker 1: book which is looking at what happened in the UK
7:28
Speaker 1: and the Industrial Revolution, So maybe say a little bit
7:31
Speaker 1: about the sort of contrasting impact of the different technologies
7:35
Speaker 1: that came in there.
7:35
Speaker 2: Thank you them for bringing that up absolutely. I mentioned
7:38
Speaker 2: already that we are so fortunate to have had the
7:43
Speaker 2: industrial technological improvement that started somewhere in the UK in
7:48
Speaker 2: the middle of the eighteenth century. We are so fortunate,
7:52
Speaker 2: but the people who live through it weren't. The first
7:54
Speaker 2: eighty ninety years of the British Industrial Revolution was dreadful
7:58
Speaker 2: for people. Incomes stagnated, working hours, expanded working conditions, Work
8:05
Speaker 2: worsened in factories which much greater discipline, much less autonomy.
8:09
Speaker 2: People were filled into unhealthy cities in which their life
8:15
Speaker 2: expectancy dropped, and there was no worker voice, no democratic process.
8:21
Speaker 2: The whole thing was just a very difficult time for
8:25
Speaker 2: most working people. But it didn't remain that way in
8:30
Speaker 2: the second half of the nineteenth century. You already see
8:33
Speaker 2: higher wages, much greater use of technology for improving conditions,
8:38
Speaker 2: both as public infrastructure, health improves and factories improved. And
8:44
Speaker 2: why why did that happen? Was that automatic? Again, our
8:47
Speaker 2: reading of history with a lot of evidence says no,
8:49
Speaker 2: that wasn't automatic. There was a complete transformation of British
8:53
Speaker 2: institutions with democracy, public sector involvement in cleaning up the cities,
9:01
Speaker 2: education and other public infrastructure, and very transformatively trade unions.
9:05
Speaker 2: You know, being a unionist was illegal in the ant
9:08
Speaker 2: Kingdom and that started changing in the second half of
9:11
Speaker 2: the nineteenth century, and that worker voice, worker negotiation were critical.
9:15
Speaker 2: As part of that process, the direction of technology changed significantly.
9:19
Speaker 2: What brought part of that misery was the automation focus
9:23
Speaker 2: and the very high discipline modern factory system. All of
9:28
Speaker 2: that started improving. No longer you could allow child labor
9:32
Speaker 2: or you know, twelve hour days and mind shaft for people,
9:36
Speaker 2: for children as young as five. All of these things
9:39
Speaker 2: were institutional in nature as well as technological.
9:42
Speaker 1: And actually I was struck by the example. I mean,
9:44
Speaker 1: we going up in Britain. You feel like you hear
9:46
Speaker 1: about the industrial revolution all the time, but I think
9:49
Speaker 1: and indeed about child labor and some of the worst aspects.
9:52
Speaker 1: But I think what was in your book you sort
9:54
Speaker 1: of bring home what a deterioration in circumstances it was
9:59
Speaker 1: that actually six seven year olds hadn't been doing twelve
10:02
Speaker 1: hours work, certainly not in the dark underground before, and
10:06
Speaker 1: so there was this a period where people's lives were
10:08
Speaker 1: actively worse. I think is worth reminding.
10:11
Speaker 2: People absolutely absolutely, and people were very exercised about it.
10:15
Speaker 2: I mean, at some point it reached such alarming proportion
10:18
Speaker 2: that middle class Brits, you know, said disc cannot go on.
10:23
Speaker 2: But all wishful thinking would not have done anything unless
10:27
Speaker 2: we changed the institutions and the direction of technology, and
10:29
Speaker 2: that's what Britain managed in the second half of the
10:31
Speaker 2: nineteenth century.
10:34
Speaker 1: So there's quite a few things there, because there's the
10:37
Speaker 1: nature of the technology and whether it tends to just
10:40
Speaker 1: replace workers or actually also produce more demand more other
10:44
Speaker 1: kinds of jobs for workers. There's also the institutions surrounding it.
10:52
Speaker 1: But part of that is about the companies that are
10:56
Speaker 1: producing the technology, that are driving technology, and how powerful
10:59
Speaker 1: they are relative to other parts of society. And I
11:02
Speaker 1: guess what the example of that is in the Gilded Age,
11:04
Speaker 1: in the US. So how does that feed into it
11:08
Speaker 1: the sort of market power of companies.
11:10
Speaker 2: I think the market power is one of the very
11:12
Speaker 2: important elements as well, because new technologies, especially those that
11:17
Speaker 2: make better use of labor, come out of the competitive process.
11:19
Speaker 2: A more diverse approach to innovation is an important part
11:23
Speaker 2: of it. Now, large companies have always been with us.
11:26
Speaker 2: I don't think we're going to be able to reverse that,
11:28
Speaker 2: and I don't think we should. Ford Motor Company started
11:32
Speaker 2: small but became one of the most important employers in
11:35
Speaker 2: the United States, and it was at the forefront of
11:38
Speaker 2: automating work. But it was also at the forefront of
11:40
Speaker 2: creating new tasks, much better working conditions for workers with
11:46
Speaker 2: higher wages, and accommodating workers into the production process so
11:50
Speaker 2: that they could actually reduce a turnover. But all of
11:54
Speaker 2: that becomes much more likely when we have countervailing powers,
11:58
Speaker 2: and countervailing powers have to have several sources. For large companies,
12:02
Speaker 2: you need competition. If they become so secure that nobody
12:06
Speaker 2: can replace them, that's not going to be good. You
12:08
Speaker 2: need countervailing powers in the form of worker voice, worker involvement.
12:12
Speaker 2: Trade unions provided that, labor movement provided that in the past,
12:15
Speaker 2: What will provide it in the future That remains to
12:17
Speaker 2: be seen, and you need the government regulation in there.
12:20
Speaker 2: You know, if companies can do whatever they want to
12:24
Speaker 2: their customers, to the environment, two workers, that's not going
12:29
Speaker 2: to lead to good outcomes. So a regulatory framework is
12:32
Speaker 2: also quite critical. Overall. I think a good way of
12:36
Speaker 2: thinking about this is democratic control of technology. Technology is
12:40
Speaker 2: something that affects us all. To say that one or
12:43
Speaker 2: two genius in Silicon Valley have to be responsible for
12:46
Speaker 2: the future of technology, I mean we all have to
12:48
Speaker 2: take whatever it's dished out to us, that's not the
12:50
Speaker 2: right perspective. And the democratic control comes from companies being
12:54
Speaker 2: at the forefront of technological progress. But those companies are
12:57
Speaker 2: threatened by rivals, countable to their workers, and they're accountable
13:01
Speaker 2: to society through democratic means.
13:03
Speaker 1: Because we do tend to think of invention and technology
13:07
Speaker 1: as a sort of as a being out being outside
13:11
Speaker 1: the system, that there's you know, people are sitting around
13:14
Speaker 1: in their labs or where you know, wherever you can,
13:17
Speaker 1: wherever you picture them, or in their garages, coming up
13:20
Speaker 1: with their ideas, and there's only it's only after a
13:23
Speaker 1: certain point that they're they're interacting with the broader world,
13:27
Speaker 1: that there is this sort of natural process of invention
13:29
Speaker 1: that happens. It doesn't feel like that process has ever
13:33
Speaker 1: been really organized or run by government or with a
13:39
Speaker 1: kind of democratic infact. So can you democratic?
13:41
Speaker 2: Interesting? You're right at some degree, But there is a
13:45
Speaker 2: broader ecosystem. First of all, a lot of innovation is
13:50
Speaker 2: coordinated by large companies. Today. If you look at the
13:55
Speaker 2: in the United States, most RND is by large and
13:58
Speaker 2: publicly traded companies. But second, even the innovation that takes
14:04
Speaker 2: place in universities, in people's garages, in small companies, it's
14:10
Speaker 2: influenced by the market system where people think profits are,
14:13
Speaker 2: and it's influenced by what we call a vision what
14:16
Speaker 2: is the best use of our scientific knowledge. And I
14:20
Speaker 2: think we have created an incorrect direction of technology because
14:25
Speaker 2: of both reasons. We have provided the wrong market incentives
14:29
Speaker 2: to digital technologies and we've provided the wrong vision. And
14:32
Speaker 2: they've both met in saying digital technologies should be designed
14:36
Speaker 2: by geniuses to be imposed on people, and they should
14:39
Speaker 2: be used for automation, for surveillance, for data collection, for
14:43
Speaker 2: reducing labour's involvement in the production process, for creating some
14:47
Speaker 2: sort of amorphous autonomous machine intelligence, and all of those
14:50
Speaker 2: are related, and they're the wrong direction. What we call
14:53
Speaker 2: for in the book is that we should strive for
14:55
Speaker 2: machine usefulness, not machine intelligence. Machines are valuable to us
14:59
Speaker 2: because they enable us to do useful things. The calculator Wikipedia,
15:03
Speaker 2: those are amazing inventions because they expand what we can do.
15:07
Speaker 2: The amorphous notion of AI that is so good that
15:10
Speaker 2: it can do everything that humans do, actually in practice
15:13
Speaker 2: not so well. But that vision, which guides a lot
15:17
Speaker 2: of research, is the wrong one.
15:19
Speaker 1: And actually, when you talk about vision, you had a
15:22
Speaker 1: fascinating phrase which for me was quite resonant in a
15:25
Speaker 1: kind of broader way, which is vision oligarchy. Tell us
15:29
Speaker 1: a bit more about that.
15:30
Speaker 2: You know, at the end of the day, I described
15:32
Speaker 2: a vision which is this machine intelligence created by a
15:40
Speaker 2: few very smart engineers and scientists that's going to transform
15:44
Speaker 2: everybody's lives. That is a very powerful vision. But where
15:49
Speaker 2: did it come from? Well, it came from Turing to
15:54
Speaker 2: some degree, but in a very different context. But it
15:58
Speaker 2: got operationalized by a number of very like minded people
16:02
Speaker 2: in Silicon Valley who've pushed this vision and have achieved
16:06
Speaker 2: some degree of commercial success early on and now are
16:11
Speaker 2: influencing the rest of society through their oversized role in
16:15
Speaker 2: the media, in all public debates, in policy, and of
16:20
Speaker 2: course their amazing wealth. And that's what we mean by
16:23
Speaker 2: division oligarchy. That's a small group of people who have
16:27
Speaker 2: captured the vision of what we can do with technology
16:30
Speaker 2: and what we should do with technology.
16:32
Speaker 1: And I guess moving on to how we think about
16:36
Speaker 1: the more recent waves of technology, you're quite damning about
16:42
Speaker 1: the impact that recent automation that AI has had. I
16:46
Speaker 1: guess some would say it's just too soon to tell
16:50
Speaker 1: how some of these technologies are really going to affect
16:52
Speaker 1: the nature of jobs and the workplace.
16:54
Speaker 2: That's right, You're right, it's too soon to tell. But
16:58
Speaker 2: there is a problem in there. It's too soon to tell.
17:00
Speaker 2: Why are we rushing to automate work so quickly? What's
17:04
Speaker 2: the rush? So? I have no doubt that automation will
17:08
Speaker 2: be part of our future. It has to be part
17:10
Speaker 2: of our future. There will be things that machines can
17:13
Speaker 2: do better than us, and nothing wrong with that, but
17:19
Speaker 2: a we should do that only when they are truly
17:22
Speaker 2: better than humans, and in humane way, rearranging work in
17:28
Speaker 2: a manner that's consistent with human priorities. And second, we
17:32
Speaker 2: should at the same time create better jobs, better tasks
17:35
Speaker 2: for humans unique skills. That's the problem. We are rushing
17:39
Speaker 2: to automate work, even when it's not so productive customer service.
17:45
Speaker 2: It's done by AI in many places, and nobody's happy
17:48
Speaker 2: with it. We've displaced workers and we are faced with
17:52
Speaker 2: these menus that are supposed to be smart. They never work.
17:57
Speaker 2: Productivity gains from that are minimal, perhaps negative. But we're
18:01
Speaker 2: rushing to do it. And at the same time we're
18:03
Speaker 2: not creating any new tasks in new jobs and new capabilities.
18:05
Speaker 2: And we can do that. AI enables us or large
18:08
Speaker 2: language models, they have the capacity to help us. As
18:13
Speaker 2: you said in your introduction, we're not doing that.
18:19
Speaker 1: I think that what you've just said describes a lot
18:22
Speaker 1: of the technologies that we've lived with over the last
18:25
Speaker 1: few years. But it does feel like CHATGPT and that
18:30
Speaker 1: much more interactive technology is different to interact with and
18:36
Speaker 1: certainly seems to be learning faster than many of these
18:39
Speaker 1: other technologies. People who are making an effort to make
18:43
Speaker 1: it part of their lives, whether it's professors or lawyers
18:48
Speaker 1: or people working in human resources are finding very quickly
18:54
Speaker 1: that it can change the way they do their work.
18:56
Speaker 1: So is there something a bit different about the generative AI.
19:00
Speaker 2: Well, I would say first, it is impressive, but part
19:05
Speaker 2: of the reason why it's impressive is because that's how
19:09
Speaker 2: it's been marketed. So what people are impressed by chat
19:14
Speaker 2: GIPT is it gives authoritative answers. It can write sonnets
19:18
Speaker 2: and poetry. It feels like these are things machines shouldn't
19:24
Speaker 2: be able to do, and that's what we're impressed by.
19:28
Speaker 2: Don't get me wrong. I do completely believe that large
19:32
Speaker 2: language models and generative AI can be used in ways
19:35
Speaker 2: that are very positive for humans, and some people have
19:39
Speaker 2: found ways of doing that with chat GPT, But chat
19:42
Speaker 2: GPT's architecture is not optimal for that. What we want,
19:47
Speaker 2: if we believe my pitch for machine usefulness, is that
19:53
Speaker 2: these programs should make us better in our jobs, in
19:57
Speaker 2: our lives, in our cognition. It doesn't work. If chachipet
20:03
Speaker 2: gives you an authoritative answer without explaining to you why
20:06
Speaker 2: you should believe it, you either believe it, which is
20:08
Speaker 2: not good, or you completely dismiss it. If I were
20:12
Speaker 2: to give you an argument, you would ask me, why
20:14
Speaker 2: are you saying that, what's your evidence? Where does that
20:17
Speaker 2: come from? Give me the provenance of that. You can't
20:20
Speaker 2: do that with chatchipt. It's not designed that way. If
20:24
Speaker 2: you ask it to provide references, it will make up some.
20:28
Speaker 2: It never really processes the reliability of information. It is
20:33
Speaker 2: not designed so that it can interact with you in
20:36
Speaker 2: a way that filters the vast amount of information that
20:39
Speaker 2: you have available. But you don't know which one is reliable.
20:42
Speaker 2: So there are many things that we could design these
20:44
Speaker 2: machines or these models differently that could be more useful
20:48
Speaker 2: to us, but as not the direction the industry is going.
20:51
Speaker 2: Many employers are excited by it not to make their
20:53
Speaker 2: labor more productive, but they want to eliminate labor.
20:56
Speaker 1: And it's interesting because I guess a lot of the
20:57
Speaker 1: commentators I was reading something by Ethan Molly other day,
21:00
Speaker 1: you know, who are excited about it, have tended to
21:04
Speaker 1: be the ones who are finding ways to make it
21:07
Speaker 1: more valuable for them. And it doesn't feel like a
21:10
Speaker 1: big leap, you know. It comes down to the way
21:13
Speaker 1: we interact with it, whether we trust its answers, whether
21:15
Speaker 1: we come back. So that doesn't seem like a big change.
21:19
Speaker 1: What you've described about changing the whole environment in which
21:23
Speaker 1: these technologies are implemented, the whole public attitude towards them.
21:28
Speaker 1: That's a pretty big change.
21:29
Speaker 2: Absolutely, absolutely, the power attitude regulation. I think these are
21:33
Speaker 2: big changes, and you're absolutely right there are people who
21:37
Speaker 2: are using chatchipitin a productive where there are some companies
21:40
Speaker 2: that have already used in productive ways. But I think
21:43
Speaker 2: the model attitude of the corporate world is not the
21:47
Speaker 2: healthy one. And that's partly because of the corporate world,
21:50
Speaker 2: but a lot also because of the way that the
21:52
Speaker 2: technology is structured right now and is marketed right now.
21:56
Speaker 1: And we should get into that because you describe certain
21:58
Speaker 1: things about it, the way it's been driven towards automation,
22:02
Speaker 1: the way that it's led down a path of being
22:07
Speaker 1: used for surveillance of individuals, and the impact that that's
22:10
Speaker 1: had on democracy. So tell us about that. How you
22:14
Speaker 1: think the technology itself has been pushed in a certain
22:19
Speaker 1: direction by its origins.
22:21
Speaker 2: Well, I think let's start with surveillance. The current field
22:25
Speaker 2: of AI is contintely intermingled with data collection, and it
22:32
Speaker 2: is hungry for data and employers are hungry for getting
22:37
Speaker 2: more information about their workers'. Governments, especially authoritarian governments, are
22:41
Speaker 2: hungry for getting more information about dissident activities. So there
22:45
Speaker 2: is a confluence of factors that is intensifying monitoring of
22:52
Speaker 2: surveillance of both citizens and workers. I think that's one
22:54
Speaker 2: of the things that we have to worry about, and
22:59
Speaker 2: generative AI is going to push more in that direction.
23:02
Speaker 2: Automation is related, but quite a distinct phenomenon. US corporations
23:10
Speaker 2: are under pressure competitively because of their shareholders, because of
23:13
Speaker 2: the vision of their managers to reduce labor costs. Nobody
23:17
Speaker 2: in the schooling system is, for example, talking about, let's
23:20
Speaker 2: hire more teachers, give them better tools, make them more skill,
23:24
Speaker 2: pay them higher wages so that they can do a
23:26
Speaker 2: better job of creating the human capital of the next generation.
23:29
Speaker 2: But that's what we need. We need much more individualized
23:31
Speaker 2: teaching in the education system. In the United States and
23:34
Speaker 2: the United Kingdom, a lot of low socioeconomic background children
23:37
Speaker 2: are having trouble getting the right type of education, the
23:41
Speaker 2: right type of skills from the schooling system. More individualized
23:46
Speaker 2: education targeted to their strengths and weaknesses could be a
23:49
Speaker 2: great boot. We can use AI for doing that, but
23:52
Speaker 2: nobody is doing that, because that means actually hiring more teachers.
23:55
Speaker 2: What schools are interested in hiring less teachers. What companies
23:58
Speaker 2: are interested in is, let's eliminate some of the more
24:01
Speaker 2: of the blue color task, let's get rid of some
24:03
Speaker 2: of the clerical tasks. So that mindset needs to change.
24:06
Speaker 1: And of course the response to that has often been, well,
24:10
Speaker 1: those individual companies, particularly in the US, those individual companies
24:13
Speaker 1: will make their decisions about how many workers they want,
24:16
Speaker 1: but the increased productivity will create more jobs in other
24:20
Speaker 1: parts of the economy. You think it's just it won't
24:23
Speaker 1: operate this time or has not always It.
24:25
Speaker 2: Will it will if it really increased productivity by a
24:28
Speaker 2: tremendous amount, it would. The question is, can we get
24:34
Speaker 2: for example, let's say three percent productivity growth in real
24:38
Speaker 2: terms every year by automating. I think that's very difficult.
24:41
Speaker 2: You're automating a few tasks in a given point in time,
24:44
Speaker 2: say even if automation is on acceleration, you're going to
24:48
Speaker 2: be automating perhaps three or four or five percent of
24:50
Speaker 2: tasks that humans do. To get that kind of huge
24:54
Speaker 2: productivity growth from automation is very difficult. That means that
24:57
Speaker 2: machines need to be ten times as productive as humans
25:01
Speaker 2: in the past. We haven't done that. In the past.
25:03
Speaker 2: We've gotten very rapid productivity growth when we made humans
25:05
Speaker 2: more productive, and I think therefore it's no surprise that
25:09
Speaker 2: today we are in a productivity slump around the world.
25:12
Speaker 2: We have six five to six times as many patterns
25:15
Speaker 2: in the United States as we did forty years ago.
25:18
Speaker 2: We have new widgets every day, amazing algorithm breakthroughs in AI,
25:23
Speaker 2: and aggregate productivity is very very anemic in the United Kingdom,
25:27
Speaker 2: it's stagnant. I think that's a course for alarm and
25:32
Speaker 2: it says that we're not using these technologies the right way.
25:35
Speaker 1: And it's so interesting because of course people look at
25:37
Speaker 1: the there is there's been a productivity slump, particularly in
25:39
Speaker 1: the UK, and often that's used as a reason why
25:43
Speaker 1: we should be accelerating our introduction of these technologies.
25:46
Speaker 2: Yes, so that's the question to me, are you going
25:48
Speaker 2: to get out of that productivity slump by doing more
25:51
Speaker 2: AI driven customer service self checkout kiosks? Is that the
25:56
Speaker 2: way to double UK productivity? I mean, you know, sure
25:59
Speaker 2: if we did self check out cuos together with better things,
26:02
Speaker 2: perhaps it could contribute.
26:04
Speaker 1: Look and as you point out in that example, that's
26:06
Speaker 1: just labour shifting because we now do the work not
26:10
Speaker 1: the casemet In banking.
26:11
Speaker 2: ATMs were introduced, but at the same time, people who
26:15
Speaker 2: used to be bank tellers became analysts, customer service reps
26:22
Speaker 2: started doing other back office tasks. So actually banking productivity
26:26
Speaker 2: increased during that period. We're not doing that latter part.
26:29
Speaker 2: We're doing ATMs on overdrive.
26:31
Speaker 1: Okay, so what do we do? How do we fix this?
26:34
Speaker 2: Well, first, we need to change the narrative. This is
26:37
Speaker 2: part of it. We need to stay away from blind
26:39
Speaker 2: techno optimism. We need to stay away from a focus
26:41
Speaker 2: on killer robots. Great in Hollywood movies, but that's not
26:44
Speaker 2: what we should be worried about. But we should be
26:47
Speaker 2: concerned about the direction of technology, and we need to
26:52
Speaker 2: center the discussion on how we can use these technologies
26:56
Speaker 2: better for democracy, better for workers, better for inequality. Then
27:00
Speaker 2: we need to start building institutions. This is not going
27:03
Speaker 2: to happen automatically, which means that we need countervailing powers.
27:06
Speaker 2: How do we have a better regulatory system? We have
27:09
Speaker 2: lost the regulatory muscle in the West. We used to
27:12
Speaker 2: regulate public utility as well We used to be able
27:15
Speaker 2: to regulate banking and financial services. Those have become harder,
27:18
Speaker 2: and we have not even tried to regulate digital technology.
27:22
Speaker 2: We need to build better democracy. Democracy has been in decline.
27:26
Speaker 2: Labor movement we need some sort of labor voice. The
27:29
Speaker 2: old model of trade unions is probably not the one
27:31
Speaker 2: for future. How do we build an organic labor movement?
27:34
Speaker 2: And then we need to talk about specific policies. Are
27:37
Speaker 2: we using the right tax tools? Are we creating the
27:41
Speaker 2: right support for private RND backed up by public RND
27:46
Speaker 2: for the right direction of technology? Is the current business
27:50
Speaker 2: model of the tech world, for example, centered on data
27:53
Speaker 2: collection and individual as digital ads? Is that the right one?
27:56
Speaker 2: Or should we actually tax digital ads? Are needs like Google, Microsoft,
28:01
Speaker 2: Facebook too big? Should we think of breaking them up?
28:04
Speaker 2: There are many policy leavers that we should be talking about.
28:07
Speaker 2: Don't claim I have the answers on those, but we
28:10
Speaker 2: make some suggestions in the book. But our purpose is
28:13
Speaker 2: not to say we know these policies will work. We
28:15
Speaker 2: need a ensemble of policies and it needs to be
28:20
Speaker 2: the result of a democratic process and expertise that brings
28:25
Speaker 2: us to the right solutions.
28:26
Speaker 1: Are there any reasons to be hopeful looking around that
28:31
Speaker 1: the government or broader society will be able to grapple
28:35
Speaker 1: with the kind of things you just talked about. I mean,
28:37
Speaker 1: we currently have in the US a Biden administration, which
28:40
Speaker 1: is by recent standards pretty activist in these directions, talks
28:44
Speaker 1: about reducing the monopoly control of the big tech companies.
28:48
Speaker 1: It's quite pro worker in the way that it's designed
28:51
Speaker 1: a lot of these big investments in it and in
28:56
Speaker 1: green technology, and yet it's realistically going to achieve a
29:00
Speaker 1: fraction of what.
29:02
Speaker 2: You've just taught me. I think the Biden administration has
29:05
Speaker 2: done very well, and indeed, as you said, it's the
29:06
Speaker 2: more most pro worker government the United States has had
29:09
Speaker 2: since FDR probably, and I upload them on passing two
29:17
Speaker 2: major policies that many people would have thought would have
29:19
Speaker 2: been impossible, the Chips Act and the IRA. But despite
29:24
Speaker 2: those high ambitions, I think they're not sufficiently focused on
29:28
Speaker 2: the direction of technology and creating the right technological environment
29:34
Speaker 2: for generating jobs for all kinds of skills. So, yes,
29:37
Speaker 2: there are many reasons to be concerned. There's only one
29:41
Speaker 2: reason for a very cautious optimism. I believe in the
29:47
Speaker 2: unique skills of humans and That's why I think automating
29:51
Speaker 2: work and surveillance are not the right direction. I think
29:53
Speaker 2: there is a diverse set of capabilities that humans have
29:58
Speaker 2: that can be very well utiliz in a new work environment.
30:04
Speaker 2: And I also believe that human ingenuity can be the
30:08
Speaker 2: best way of furthering our productivity growth. So I hope
30:14
Speaker 2: that those great opportunities are used.
30:18
Speaker 1: Final question. You're an economic historian looking back in that
30:23
Speaker 1: history that you describe with Simon Johnson in the book.
30:26
Speaker 1: One thing that is clear is although the right kind
30:30
Speaker 1: of institutions to make this technology work for people did appear,
30:35
Speaker 1: they came much more slowly than the technology itself, and
30:39
Speaker 1: that technology was not coming nearly as fast as what
30:43
Speaker 1: we're seeing now. When you look realistically at the history,
30:47
Speaker 1: does it not take an enormous amount of negative impact
30:50
Speaker 1: for there to be a response from society to make
30:56
Speaker 1: this work better? Aren't we going to have to live
30:58
Speaker 1: through quite a lot of bad things?
31:00
Speaker 2: Great question, and you know that's something that worries me.
31:04
Speaker 2: We don't actually discuss in the book, because it has
31:07
Speaker 2: gelled in my mind more recently. Early industrial revolution created
31:13
Speaker 2: a lot of misery, as we talked about, and there
31:16
Speaker 2: was nothing to be bilittled about that. But when reforms
31:23
Speaker 2: and policy responses and the labor movement's reactions came, it
31:27
Speaker 2: wasn't too late and things could be reorganized. Things are
31:30
Speaker 2: happening much faster today. Are we going to be too late,
31:34
Speaker 2: if not today, in the next month or next year.
31:38
Speaker 2: I don't know the answer to that, but I was
31:40
Speaker 2: worried enough that I was one of the early signatories
31:43
Speaker 2: for the letter that asked for a six month pose
31:45
Speaker 2: on the training of large language models. Not because I
31:48
Speaker 2: agreed with the text. There was a lot of stuff
31:51
Speaker 2: that are about super intelligent AI that I definitely don't
31:55
Speaker 2: worry about. That's not the top of my agenda. But
31:59
Speaker 2: I thought that building a broad coalition of academic and
32:02
Speaker 2: entrepreneurial voices to say, let's just take some time. The
32:09
Speaker 2: loss to humanity if we are six months late in
32:12
Speaker 2: implementing some AI technology is trivial. The damage we can
32:17
Speaker 2: do by irreversibly destroying democracy or cementing an approach that's
32:25
Speaker 2: not the right one could be much much larger. So
32:28
Speaker 2: take some time. There's no rush here.
32:31
Speaker 1: But I guess the other lesson of history is there
32:33
Speaker 1: are some things that are unstoppable.
32:35
Speaker 2: I don't think technology's direction is preordained, and sure technology
32:44
Speaker 2: should not be stopped, and in some sense advances are unstoppable.
32:48
Speaker 2: But we can choose its direction, and doing so deliberatively
32:52
Speaker 2: building the right institutions is feasible.
32:54
Speaker 1: Darness Mark, thank you so much.
32:56
Speaker 2: Thank you, it was a true pleasure to be here.
33:05
Speaker 1: That's it for this special episode of Stephanomics. We'll be
33:07
Speaker 1: back next week. In the meantime, you can, as always,
33:10
Speaker 1: get a lot more economic insight and news from the
33:13
Speaker 1: Bloomberg Terminal website or app. This episode was produced by Samasadi,
33:18
Speaker 1: with special thanks to darn Asmo Blue and Ruth Kick.
33:21
Speaker 1: The executive producer of Stephanomics is Molly Smith and the
33:24
Speaker 1: head of Bloomberg Podcast is Sage Bowman.