Latest / kill switch / Dull, Dirty, Dangerous
Transcript
- 0:04Speaker 1: Sleepwalkers is a production of I Heart Radio and unusual productions.
- 0:12Speaker 1: AI will make phenomenal companies and tycoons faster, and it
- 0:18Speaker 1: will also displace jobs faster than computers and the Internet.
- 0:23Speaker 1: It's already happening. That's Kai Fu Lee speaking, the former
- 0:27Speaker 1: head of Google China and the so called oracle of AI.
- 0:30Speaker 1: I think there are at least two issues involved. One
- 0:34Speaker 1: is how to do income redistribution, and that is a
- 0:39Speaker 1: very complex issue. I'm not an expert, but one way
- 0:42Speaker 1: or another, the ultra rich who did extremely well based
- 0:47Speaker 1: on AI or other reasons, I think somehow need to
- 0:50Speaker 1: help the people who are under privileged or even victimized
- 0:54Speaker 1: by technology. The exact mechanism I don't know, but if
- 0:58Speaker 1: we don't do it, redistribution is going to be a
- 1:01Speaker 1: serious matter for our social stability. It's not actually a
- 1:05Speaker 1: underprivileged minority, it will become an underprivileged majority. The benefits
- 1:11Speaker 1: of the AI revolution will not be evenly distributed, and
- 1:14Speaker 1: according to Kaifu, automation will replace fort of jobs worldwide
- 1:19Speaker 1: in the next fifteen years. The second part is how
- 1:25Speaker 1: do we help people whose jobs have been displaced find
- 1:28Speaker 1: the new beginning? We ask the question what can AI
- 1:32Speaker 1: and automation not do? That is the central question this episode,
- 1:37Speaker 1: as AI and Automation displays more and more jobs, What
- 1:40Speaker 1: will be left for us to do and who will
- 1:43Speaker 1: be qualified to do it? Today will explore the automated
- 1:46Speaker 1: economy and the changes it will bring. Im as Velashen
- 1:49Speaker 1: Welcome to Sleepwalkers. So, Carol, when I hear Kaifou talking
- 2:05Speaker 1: about jobs being lost to AI, my mind goes immediately
- 2:08Speaker 1: to driverless cars and self driving cars replacing taxis, um,
- 2:13Speaker 1: long distance trucking, that kind of thing. Yeah, but there's
- 2:15Speaker 1: also you know, agriculture, like combine harvests, like robots who
- 2:20Speaker 1: are picking fruit. Um. Washington State actually announced that next
- 2:24Speaker 1: season they're going to be rolling out these vacuum harvesters
- 2:26Speaker 1: that use AI to identify and pick only ripe apples. Wow,
- 2:30Speaker 1: so not only picking the fruit, but also being smart
- 2:32Speaker 1: about which fruit it picks. That's right, the ripe stuff,
- 2:36Speaker 1: the stuff. And there's actually this raspberry picking robot in
- 2:40Speaker 1: the UK that was funded by some British supermarkets and
- 2:45Speaker 1: those robots can pick twenty five thousand berries a day
- 2:49Speaker 1: versus a humans fifteen thousand in an eight hour day,
- 2:52Speaker 1: and also remember this, eight hour days for human being
- 2:56Speaker 1: is a long day for a robot. A robot doesn't
- 2:58Speaker 1: know what a long day is, nor does it know
- 2:59Speaker 1: what a short day is. And it can work into
- 3:01Speaker 1: the night right and when we force ourselves into comparison
- 3:03Speaker 1: with these robots, that kind of creates very realistic expectations
- 3:08Speaker 1: for workers can do. Interesting is not just jobs that
- 3:11Speaker 1: require mechanical skills that Kaifu thinks will be lost to automation.
- 3:15Speaker 1: And AI actually doesn't distinguish between white collar and blue
- 3:19Speaker 1: collar jobs. So any job that has a routine element,
- 3:22Speaker 1: whether it's underwriting loans or telemarketing or researching, you know,
- 3:26Speaker 1: this is a lot of work. The first AI podcast
- 3:28Speaker 1: may not be too far off. Um. It actually reminds
- 3:31Speaker 1: me the episode we did about AI and creativity that
- 3:34Speaker 1: algorithms that can write poetry and music and screenplays are
- 3:37Speaker 1: already here. This is not some robot apocalypse in the
- 3:41Speaker 1: distant future. Job displacement is with us. Julian, You've got
- 3:44Speaker 1: in touch with somebody who's seeing this play out in
- 3:46Speaker 1: real time. Yeah, it did. His name's Wild Kankowski and
- 3:50Speaker 1: he lives in Florida, all around the city whatever direction
- 3:54Speaker 1: we're gonna go. We know where every every McDonald's pretty
- 3:59Speaker 1: much is on the did a job well. A lot
- 4:01Speaker 1: of the people know us because we go in there
- 4:04Speaker 1: all the time. A lot of them know me because
- 4:07Speaker 1: not too many people get a medium coffee with twelve creams. Yeah,
- 4:15Speaker 1: as what is taking a huge number of creams and
- 4:17Speaker 1: his coffee? What? He owns the pool screens and repair
- 4:20Speaker 1: business in Orlando, Florida. His job takes him all around town,
- 4:23Speaker 1: but every morning starts the same way Adam McDonald's and
- 4:27Speaker 1: recently what he has seen a change. They just started
- 4:31Speaker 1: to show up, probably about a year or so ago.
- 4:34Speaker 1: That way, when we go to a counter, people are
- 4:38Speaker 1: getting mad because they want you to go to use
- 4:42Speaker 1: the key off and I'm walking up to the counter
- 4:44Speaker 1: wanting to get my coffee and get on on our day.
- 4:48Speaker 1: They're like, oh, you got to use the kios so
- 4:50Speaker 1: and then they want me to hit the screen. The
- 4:52Speaker 1: screen says, go to this thing, go to beverage. Okay,
- 4:57Speaker 1: what kind of beverage? Well, okay, go the coffee, but
- 5:01Speaker 1: what do you want? Ice coffee? This? That? And then
- 5:03Speaker 1: instead of me saying twelve cream and she hears me.
- 5:07Speaker 1: Now I get to hit the machine like twelve times
- 5:09Speaker 1: that that that that that that that that that that
- 5:11Speaker 1: twelve times to get it, because that's how many times
- 5:14Speaker 1: I get to hit it to get to twelve. The
- 5:16Speaker 1: thing is knocking someone out of a job. We've all
- 5:19Speaker 1: been wally stuck at a self checkout or yelling at
- 5:22Speaker 1: an automated phone menu that refuses to understand what we're saying.
- 5:26Speaker 1: But those interactions are not just frustrating for us. They're
- 5:29Speaker 1: real world examples of jobs being displaced by technology, and
- 5:33Speaker 1: they don't only affect the people whose jobs are threatened.
- 5:36Speaker 1: We're in a lot of different McDonald's and I probably
- 5:40Speaker 1: recognize every single person in there. Some people I've known
- 5:44Speaker 1: probably ten fifteen years, and they know who I am.
- 5:48Speaker 1: You know, they're friendly enough to make you feel a
- 5:51Speaker 1: little special there. That way, I guess we might be
- 5:55Speaker 1: walking through a store and then I'll see those people
- 5:58Speaker 1: and I'll go over and them say, yeah, you're for
- 6:00Speaker 1: McDonald's or that, and then they'll be like, yeah, I
- 6:02Speaker 1: know who you are. Then you actually get them meet
- 6:05Speaker 1: and greet someone and make a conversation for a minute
- 6:08Speaker 1: or two. That way, why would human contact when you
- 6:12Speaker 1: talking to a person for a second and getting my
- 6:15Speaker 1: food and paying them in another two seconds. There shouldn't
- 6:19Speaker 1: have been nothing wrong with that process. So, Julie, how
- 6:23Speaker 1: did this come about? What made you want to include
- 6:25Speaker 1: Wally story in the podcast? Well, for one thing, I
- 6:27Speaker 1: love Wally, but these are also familiar stories, right, I mean,
- 6:31Speaker 1: and while he's been able to see this one play
- 6:33Speaker 1: out over time, where you can see how just changing
- 6:35Speaker 1: one part of one task the way he orders a
- 6:38Speaker 1: coffee has actually had this ripple effect that also follows
- 6:41Speaker 1: him around as he goes about his day. Yeah. I
- 6:43Speaker 1: was especially struck by Wally story because it's easy to
- 6:46Speaker 1: talk about automation and job displacement as these big abstract ideas,
- 6:50Speaker 1: but here's somebody who's actually felt it. Even though it's
- 6:53Speaker 1: not his job that's been lost. Is something that affects
- 6:56Speaker 1: the whole community. You know, I don't mean to be
- 6:59Speaker 1: super nestar algic, but a lot of great movies and
- 7:03Speaker 1: great young adult novels have you know, the teenage girl
- 7:08Speaker 1: who's angsty and you know works at the Friar and
- 7:12Speaker 1: you know, now it's just like you're gonna have like
- 7:13Speaker 1: an angsty data scientist, you know, mulling over the express
- 7:17Speaker 1: checkout crouched over the screen. Well, those those golden arches,
- 7:21Speaker 1: they are very enduring symbol for America UM And earlier
- 7:25Speaker 1: this year, McDonald's acquired an AI company for three hundred
- 7:28Speaker 1: million dollars. It was their biggest acquisition for twenty years.
- 7:32Speaker 1: And it's all about predicting what people might order before
- 7:35Speaker 1: they even arrive at the store. So even the days
- 7:37Speaker 1: of Kiosks maybe numbered, maybe we'll be nostalgic about them
- 7:40Speaker 1: in twenty years, but nonetheless, this AI acquisition could ultimately
- 7:44Speaker 1: lead to a better customer experience. And it's important to
- 7:47Speaker 1: remember that the AI revolution doesn't need to be just
- 7:50Speaker 1: about displacing jobs. It can also be about augmenting us
- 7:54Speaker 1: and our experience. One person working on human machine partnership
- 7:58Speaker 1: is Gil Pratt see EEO of the Toyota Research Institute.
- 8:03Speaker 1: Many of our colleagues at other companies are really focused
- 8:07Speaker 1: on building only the self driving car, where you replace
- 8:11Speaker 1: the driver with an AI system. But we also have
- 8:14Speaker 1: this other track of building something that we call the Guardian,
- 8:17Speaker 1: which is meant to safeguard a human being when they drive,
- 8:20Speaker 1: to avoid accidents and to avoid crashes. I think the
- 8:24Speaker 1: Guardian approach has been at odds because of money. The
- 8:28Speaker 1: economic desire to replace the driver in a taxi is
- 8:33Speaker 1: very large, and a lot of companies are sort of
- 8:35Speaker 1: going after this attractive idea of automating out the human
- 8:39Speaker 1: being from driving taxis. But you know, Toyota is first
- 8:42Speaker 1: and foremost a car company, which means that we have
- 8:46Speaker 1: this business of making cars. We also want to make
- 8:49Speaker 1: cars a lot more safe, and we also want to
- 8:51Speaker 1: make them a lot more fun. Gil makes an important
- 8:55Speaker 1: point today, our innovation is driven by the market. Companies
- 8:59Speaker 1: like Uber in tested to keep their valuations high by
- 9:02Speaker 1: promising their investors that they will be able to do
- 9:04Speaker 1: better business in future by replacing human drivers. Toyota is
- 9:08Speaker 1: actually an investor in Uber, but it's primary business is
- 9:11Speaker 1: car manufacturing, so there that is on enhancing the abilities
- 9:15Speaker 1: of human drivers rather than replacing them, making driving more
- 9:19Speaker 1: fun and Gil's humanistic approach to technology is also being
- 9:23Speaker 1: applied to other problems at the Toyota Research Institute. We
- 9:26Speaker 1: want to allow people to age in place with dignity,
- 9:31Speaker 1: and in particular, we want to help them by amplifying
- 9:36Speaker 1: their abilities to make up for what was lost, rather
- 9:39Speaker 1: than replacing their abilities. And make them feel as if
- 9:43Speaker 1: they're elderly. It's a subtle difference, and it's very easy
- 9:47Speaker 1: to get it wrong. It's very easy to build a
- 9:51Speaker 1: technology that is ostensibly going to help some someone, but
- 9:56Speaker 1: it's what it's really doing is offloading work from them
- 9:59Speaker 1: and making them feel they can't do it and therefore
- 10:01Speaker 1: they're old and they should just sit in a chair.
- 10:04Speaker 1: It's much harder to figure out a way, particularly in
- 10:08Speaker 1: the robotics field, to continue to engage the person so
- 10:12Speaker 1: that they feel like they can do it themselves. And
- 10:16Speaker 1: that's a little bit of a difference in how we
- 10:17Speaker 1: try to do things. There's one that we've recently started
- 10:21Speaker 1: to show, which is a machine called the Buddy, and
- 10:25Speaker 1: this idea is one where older people have a lot
- 10:28Speaker 1: of difficulty reaching down low to pick up things from
- 10:31Speaker 1: the ground and difficulty moving heavy things, and so we're
- 10:36Speaker 1: working on a machine that still has the human in
- 10:39Speaker 1: the loop, but makes it much easier for them to
- 10:42Speaker 1: do that task. But it understands that no matter how
- 10:46Speaker 1: much robotics may be able to help solve the practical
- 10:49Speaker 1: challenges of life as an older person, it can never
- 10:52Speaker 1: replace a human cab provider. Just to be very very clear,
- 10:56Speaker 1: we don't want to replace people as companions. We think
- 11:00Speaker 1: it what human beings want most of all in a
- 11:03Speaker 1: companion is another human being, so companion. This brings us
- 11:10Speaker 1: back to what Kaifu was saying right at the top
- 11:12Speaker 1: of the episode, what can AI and automation not do?
- 11:16Speaker 1: So Yeah, Gil acknowledges that no matter how much progress
- 11:19Speaker 1: is made in the field of robotics to help elderly people,
- 11:22Speaker 1: nothing's going to make up for human contact. I actually
- 11:25Speaker 1: was able to talk to Sherry Turkle, who's a professor
- 11:28Speaker 1: at m i T who talks a lot about human
- 11:30Speaker 1: beings and their relationship with technology, and she talks about
- 11:34Speaker 1: this fluffy seal robot called Pero, which is used in
- 11:37Speaker 1: nursing homes to soothe Alzheimer's patients. And it can simulate
- 11:42Speaker 1: this like affectionate little animal, and it can be really
- 11:44Speaker 1: effective at drawing people out of their shells when they're
- 11:47Speaker 1: otherwise hard to reach or feeling disoriented. On the other hand,
- 11:50Speaker 1: and this is Sherry's argument, it becomes really easy for
- 11:53Speaker 1: family members to be like, well, you know, my grandpa
- 11:55Speaker 1: has this, you know, seal at home. I don't need
- 11:57Speaker 1: to go visit him all the time. And I know
- 11:59Speaker 1: that sounds extreme, but it's more of the idea of
- 12:01Speaker 1: the fact that we're using these robots to make us
- 12:05Speaker 1: feel better about calming people who we could otherwise have
- 12:09Speaker 1: strong relationships with. Yeah, and I think it also normalizes
- 12:12Speaker 1: the idea of interacting with robots or technology instead of
- 12:16Speaker 1: real people. And that's painful. That's what wal he was
- 12:19Speaker 1: really talking about. Yes, it's frustrating to have to use
- 12:21Speaker 1: the Kiosk when you want twelve creems with your coffee.
- 12:24Speaker 1: But more important, Leader Rhodes Community bonds. It's no wonder
- 12:28Speaker 1: that a company like McDonald's is spending a ton of
- 12:31Speaker 1: money on this. It makes them more efficient and profitable
- 12:35Speaker 1: if they don't have to pay people. Yeah, and it's
- 12:37Speaker 1: hard to turn back the clocks. You know. Donald Trump
- 12:39Speaker 1: talks about bringing back the cold jobs, but jobs that
- 12:42Speaker 1: have been lost are very hard to recreate. It doesn't
- 12:45Speaker 1: make me think about Kaifu's comment at the top of
- 12:47Speaker 1: the episode about the underprivileged majority. Uvell know Harari, who's
- 12:51Speaker 1: coming to join us later in the series, talks about
- 12:54Speaker 1: a useless class. When we come back, we look at
- 12:57Speaker 1: what this means for the people at the sharp end
- 12:59Speaker 1: the people losing their jobs to automation, and at some
- 13:02Speaker 1: of the proposed solutions. According to an OXFAM International report
- 13:12Speaker 1: published earlier this year, the twenty six richest billionaires in
- 13:16Speaker 1: the world have as much wealth as the poorest three
- 13:19Speaker 1: point eight billion people, and many of those billionaires made
- 13:22Speaker 1: their fortunes from technology. Jeff Bezos is the world's richest
- 13:26Speaker 1: person thanks to Amazon. Meanwhile, Amazon is investing hundreds of
- 13:30Speaker 1: millions of dollars in automating their supply chain, in other words,
- 13:34Speaker 1: attempting to cut out the labor force who made the
- 13:36Speaker 1: business possible. It's a bit like Uber's investment in self
- 13:39Speaker 1: driving technology. So what jobs might be safe from the
- 13:42Speaker 1: relentless march towards automation, I asked Kai Fu Lee. My
- 13:47Speaker 1: general feeling is that these will be the human interaction jobs,
- 13:50Speaker 1: the compassion and empathetic jobs, the jobs that we expect
- 13:54Speaker 1: a human and refused to work with a robot. That
- 13:57Speaker 1: would doubly ensure these jobs are safe as one AI
- 14:01Speaker 1: can't do them now, and too, even if AI got better,
- 14:04Speaker 1: customers don't accept it, then those jobs will become the
- 14:08Speaker 1: right areas to retrain people to move into so jobs
- 14:11Speaker 1: like nurses, Nanni's elderly care, high end jobs like psychiatrists
- 14:18Speaker 1: and doctors, because the future it will be different. AI
- 14:21Speaker 1: can do the analytical part, but the doctor will still
- 14:24Speaker 1: need to provide the warmth and the human contact that
- 14:27Speaker 1: the patient expects during the worst period of vulnerability. What
- 14:33Speaker 1: we may move more towards ordering from kiosks and help menus,
- 14:37Speaker 1: or not even needing to order at all. Kaifu agrees
- 14:40Speaker 1: with Gil, will still need to human touch in a
- 14:42Speaker 1: range of industries, many of them senters around care and
- 14:45Speaker 1: human services, and it's striking to hear these two pioneers
- 14:49Speaker 1: of new technology. Kaifu and AI and Gil in robotics agree,
- 14:54Speaker 1: both arguing that automation might increase the value of what
- 14:57Speaker 1: is uniquely human. Guilt terms the history. To back up
- 15:01Speaker 1: his argument, he looks at how our understanding of our
- 15:03Speaker 1: own value as humans shifted during the Industrial Revolution away
- 15:08Speaker 1: from the ability of our bodies towards the ability of
- 15:11Speaker 1: our minds. You know, if you go back in history
- 15:14Speaker 1: and you say, how did people earn a living back
- 15:17Speaker 1: in the days of mechanical work, There wasn't you know,
- 15:20Speaker 1: steam engines, no use of gasoline or oil or anything
- 15:23Speaker 1: like that. And the answer was that the economic capital
- 15:27Speaker 1: that a human being would have just by being born
- 15:30Speaker 1: was primarily mechanical. So our muscles made us worthwhile at
- 15:34Speaker 1: a minimum level, and machines effectively took over most of
- 15:37Speaker 1: the mechanical work that we do, and so we now
- 15:40Speaker 1: are valued mostly what we can do with our minds,
- 15:44Speaker 1: assuming that this next stage of AI occurs where most
- 15:48Speaker 1: of the mental labor that is done is displaced. What
- 15:52Speaker 1: I think we need to think about now is what
- 15:55Speaker 1: will we do then? And we need to think about
- 15:57Speaker 1: it even if this next stage of A doesn't come
- 16:00Speaker 1: for a while, because we went from mechanical to mental.
- 16:04Speaker 1: Is there something next? Is there something next? That is
- 16:09Speaker 1: the trillion dollar question? According to Guilty, industrial revolution led
- 16:13Speaker 1: us to place more value on the mind than the muscle.
- 16:16Speaker 1: Now that a I can increasingly perform mental labor, but
- 16:19Speaker 1: we find a new source of value. And could it
- 16:21Speaker 1: be like CAIFU hinted at as well, some emotional connection.
- 16:26Speaker 1: When I read a story to my son, it matters
- 16:29Speaker 1: a whole lot to him. When I read a story
- 16:32Speaker 1: to my mother, it's very much the same thing. So
- 16:35Speaker 1: could we actually decide to increase the value that we
- 16:39Speaker 1: pay for social work. There's many, many different jobs that
- 16:42Speaker 1: really should be paid much much higher than they are now,
- 16:45Speaker 1: jobs of teaching and helping so forth. And so I'm
- 16:49Speaker 1: an optimist that we can find an answer, but I
- 16:51Speaker 1: think we need to realize the difficulty in order to
- 16:54Speaker 1: move towards that answer. The difficulty is huge because as
- 16:58Speaker 1: of now, excepting the luxury, the market does not reward
- 17:02Speaker 1: the kind of human contact that Kaifu and Gil allude to.
- 17:06Speaker 1: And while we, like Wally may wish for our food
- 17:08Speaker 1: orders not to be automated, how much more would we
- 17:11Speaker 1: pay for human contact? How much more could we afford
- 17:14Speaker 1: to pay. Part of the problem is that automation is
- 17:17Speaker 1: exacerbating the gap between rich and poor. Technology companies can
- 17:21Speaker 1: increasingly create wealth without needing to pay the wages of
- 17:24Speaker 1: additional employees. That's the secret behind that word you hear
- 17:28Speaker 1: so often scale, which is why Kai Fu Lee proposes
- 17:32Speaker 1: a radical solution. If we start to redistribute the income
- 17:36Speaker 1: that is taking away the power of the ultra ridge,
- 17:39Speaker 1: If we start to give the people who are stripped
- 17:43Speaker 1: of their current jobs a new job that has not
- 17:45Speaker 1: only income but also meaning, I think um people would
- 17:50Speaker 1: be more fulfilled, their children at least would have a
- 17:52Speaker 1: chance just to pause. Kai Foo Lee is a hugely
- 17:55Speaker 1: successful international investor arguing that we need to overturn and
- 18:00Speaker 1: one of the most fundamental assumptions of American society that
- 18:04Speaker 1: the market should be allowed to set the price. And
- 18:06Speaker 1: Kaifu is not alone. Others in Silicon Valley are calling
- 18:10Speaker 1: for a so called universal basic income as SI pen
- 18:13Speaker 1: pay to all citizens to acknowledge an increasingly broken relationship
- 18:17Speaker 1: between labor and value. Today, we're nowhere close on either
- 18:21Speaker 1: of those ideas, but a growing course of inside voices
- 18:24Speaker 1: is acknowledging that automation will bring further disruption to society,
- 18:29Speaker 1: and others have even greater fears. You may remember Ian
- 18:32Speaker 1: Bremer from our episode on China and Surveillance. He's a
- 18:35Speaker 1: political scientist and the author of Us Versus Them, The
- 18:39Speaker 1: Failure of Globalism. I am less worried about just jobs
- 18:45Speaker 1: going away, then I am about technology facilitating the creation
- 18:51Speaker 1: of completely different types of human beings. What happens when
- 18:56Speaker 1: you have the ability to actually provide comp copletely different
- 19:00Speaker 1: sets of cognitive skills to human beings that have access
- 19:04Speaker 1: to certain types of new technology ian sphere is that
- 19:09Speaker 1: as technology improves, the rituals simply reproduced their privilege through
- 19:13Speaker 1: elite universities and professional networks, they may start to upgrade
- 19:17Speaker 1: their very hardware, making social mobility even harder for those
- 19:21Speaker 1: who can't afford the same modifications. Better memory retention, better
- 19:26Speaker 1: pattern recognition, more ability to link to real time information,
- 19:31Speaker 1: and the global net I mean, ability not to sleep
- 19:34Speaker 1: for longer periods of time, all of this sort of thing. Right,
- 19:38Speaker 1: The danger is that I don't care how much money,
- 19:41Speaker 1: how much wealth in society, and when you start creating
- 19:44Speaker 1: that kind of differentiation, everything we know about human history
- 19:48Speaker 1: is that that doesn't end well. Those other people that
- 19:52Speaker 1: aren't as capable get treated like animals or worse. And
- 19:56Speaker 1: I am very deeply worried that the speed of technol
- 20:00Speaker 1: logical transformation, coupled with the speed of this new industrial revolution,
- 20:04Speaker 1: makes it much more likely that large numbers of people
- 20:08Speaker 1: in our own societies, not in other countries, but like
- 20:11Speaker 1: right here, are suddenly not going to have that capacity,
- 20:15Speaker 1: and we're going to treat them as different types of humans,
- 20:19Speaker 1: maybe not even as humans at all. This is the
- 20:22Speaker 1: truly dystopient future that we will fear carat this concept
- 20:25Speaker 1: of a two track humanity facilitated by technology, where some
- 20:29Speaker 1: people have value and others don't. Yeah, you know, this
- 20:32Speaker 1: is the dark version of trans humanism, which we're going
- 20:35Speaker 1: to talk about later in the series. But you know
- 20:37Speaker 1: it's not some sci fi fantasy. Our favorite pre super
- 20:42Speaker 1: villain Elon Musk, founded Neuralink, which aims to create brain
- 20:46Speaker 1: computer interfaces. Like why do we need that? Well, I
- 20:50Speaker 1: guess because in today's economy, being smart is seen as
- 20:53Speaker 1: the most important differentiating factor. But we're not talking about
- 20:56Speaker 1: being an intellectual, like, we're talking about being cognitively enhanced
- 20:59Speaker 1: by a computer or by technology. And Elon Musk isn't
- 21:03Speaker 1: the only person who's noticed how important it is to
- 21:05Speaker 1: be cognitively enhanced, shall we say? Last year, the World
- 21:08Speaker 1: Bank announced the program called the Famine Action Mechanism to
- 21:12Speaker 1: get relief to famine hit areas faster, and they explicitly
- 21:15Speaker 1: said one of the reasons they're doing this is that
- 21:17Speaker 1: because people who are malnourished in the womb may have
- 21:20Speaker 1: cognitive issues later in life and thus be unable to
- 21:23Speaker 1: compete in the new economy. You know, I found it
- 21:25Speaker 1: really interesting that this program is actually powered by AI.
- 21:29Speaker 1: It draws on data like social media, food prices, rainfall,
- 21:33Speaker 1: and then automatically assigns funds so that money gets where
- 21:37Speaker 1: it's needed before it's too late. It's a textbook case
- 21:40Speaker 1: of what AI can do and we can't, which is
- 21:41Speaker 1: to notice these patterns and correlations between different types of
- 21:44Speaker 1: data sets which are so big as to be impossible
- 21:47Speaker 1: for us to compute, and as so often in Sleepwalkers,
- 21:50Speaker 1: and it's an example of technology being a double edged short.
- 21:54Speaker 1: On the one hand, it may be widening the gap
- 21:55Speaker 1: between rich and poor, but on the other hand, it
- 21:58Speaker 1: can potentially feed the world. When we come back, we
- 22:01Speaker 1: explore other ways AI and robotics can revolutionize food production.
- 22:12Speaker 1: We've looked at how AI and robotics could exacerbate the
- 22:15Speaker 1: gulf between rich and poor, and how this new industrial
- 22:18Speaker 1: revolution could put a new value on human connection. But
- 22:22Speaker 1: could we use automation to actually decrease global inequality? One
- 22:27Speaker 1: key factor is access to quality nutrition and roboticist George
- 22:31Speaker 1: Kantor gave a talk last year at south By Southwest
- 22:34Speaker 1: called AI will help feed a growing planet. I wanted
- 22:38Speaker 1: to learn more, so I called him for a conversation
- 22:40Speaker 1: from his office at the Robotics Institute of Carnegie Mellon.
- 22:44Speaker 1: A lot of people when they think about robots and
- 22:47Speaker 1: technology being used to assist agriculture, think about robots driving
- 22:52Speaker 1: around and picking grapes or plowing fields and things like that.
- 22:55Speaker 1: But despite being a robotics expert, George is currently focusing
- 22:58Speaker 1: on crop genetics. The way plant breeding works, you have
- 23:01Speaker 1: a bunch of parents. Uh plant breeder very carefully uses
- 23:06Speaker 1: all his or her experience to figure out which parents
- 23:08Speaker 1: will make the best potential children. They make those crosses.
- 23:13Speaker 1: They then do these field trials where they grow the
- 23:15Speaker 1: child varieties and they measure them and see how they do,
- 23:19Speaker 1: and then the winners go back in the pool and
- 23:21Speaker 1: the losers they weed out. One of the crops we
- 23:24Speaker 1: work with is sorghum. It's grown all over the world.
- 23:27Speaker 1: They're like forty different varieties of it. In particular, the
- 23:30Speaker 1: grain sorghum variety is a staple crop in places like
- 23:33Speaker 1: Sub Saharan Africa and India, parts of the world where
- 23:37Speaker 1: population is growing more rapidly than the rest of the
- 23:39Speaker 1: planet's populations, and the predictions for the impact of global
- 23:42Speaker 1: warming are are pretty high. Jewles uses technology to make
- 23:45Speaker 1: the work if human plant breed is dramatically more efficient.
- 23:49Speaker 1: But this work is completely invisible to consumers. So we
- 23:52Speaker 1: have built a robot that goes out to a breeding
- 23:54Speaker 1: experiment where a breeder has grown a thousand different varieties
- 23:57Speaker 1: of sorghum are robot goes through and takes all these
- 24:00Speaker 1: detailed measurements about how the plants are growing throughout the year,
- 24:03Speaker 1: and then the breeder can use those measurements to make
- 24:05Speaker 1: better decisions. The end user of this process I'm describing
- 24:09Speaker 1: won't see any technology at all. They will get a
- 24:12Speaker 1: seed that looks just like the seed they get now,
- 24:14Speaker 1: except it will be a little bit better because the
- 24:16Speaker 1: breeder improved it using our robots. These invisible changes to
- 24:21Speaker 1: the food production system can have huge consequences. Better seeds
- 24:25Speaker 1: mean better yields and could ultimately lead to a better
- 24:28Speaker 1: nourish world. But George isn't only thinking about how to
- 24:31Speaker 1: make heartier, better plants. He's also thinking about another problem,
- 24:34Speaker 1: how will we efficiently feeded global population who increasingly live
- 24:38Speaker 1: in cities and not on the farm. Imagine every building
- 24:42Speaker 1: in a city had a little greenhouse hanging off the
- 24:45Speaker 1: side of it, or a little growing room in the basement,
- 24:48Speaker 1: and now you've got these indoor growing systems that tend
- 24:50Speaker 1: to like generate more heat than they need, so one
- 24:53Speaker 1: of their big problems is venting off the heat. Well,
- 24:56Speaker 1: buildings have to pay a lot of money to heat
- 24:58Speaker 1: the buildings. So if you had this sort of sim
- 25:00Speaker 1: the artic relationship between the people in the building and
- 25:02Speaker 1: the plants in the building, they can exchange heat, and
- 25:04Speaker 1: they can exchange atmosphere and all kinds of things. If
- 25:07Speaker 1: you take that idea and you scale it up to
- 25:09Speaker 1: like a city scale, where now you have dozens or
- 25:11Speaker 1: hundreds of buildings that all have these different energy needs
- 25:14Speaker 1: and different agricultural needs, and they're all sort of sharing.
- 25:18Speaker 1: You have some sort of overarching AI that controls what
- 25:22Speaker 1: energy gets moved where. Um, you can imagine that there
- 25:25Speaker 1: are big efficiencies that can be gained. George's outlining a
- 25:29Speaker 1: vision where robotics and AI help us tackle one of
- 25:32Speaker 1: the world's most enduring sources of inequality food access, and
- 25:36Speaker 1: doing so could also make agriculture more energy efficient and
- 25:40Speaker 1: thus begin to address another huge problem that will disproportionately
- 25:43Speaker 1: affect the world's poorest people, climate change. So yes, automation
- 25:48Speaker 1: will take jobs away, but it can also potentially raise
- 25:51Speaker 1: quality of life and the quality of the global environment.
- 25:55Speaker 1: And as far as George is concerned, the type of
- 25:58Speaker 1: labor being replaced is not exact the work that maximizes
- 26:01Speaker 1: human potential. We call them dull, dirty, dangerous, so jobs
- 26:05Speaker 1: that people don't want or are dangerous to do, or
- 26:08Speaker 1: people are getting injured in. When I go visit the
- 26:11Speaker 1: Great industry in California and I see the laborers, they're
- 26:15Speaker 1: out there, they're stooped over under trees, They're doing this
- 26:18Speaker 1: extremely backbreaking labor. There are high incidences of repetitive stress injuries,
- 26:24Speaker 1: and so it's just not a very pleasant environment to
- 26:27Speaker 1: be working in. When automation comes into an industry, it
- 26:31Speaker 1: takes away some jobs that were there, but it creates
- 26:35Speaker 1: other opportunities. So for example, most orchards, you know, they'll
- 26:39Speaker 1: have sort of a year round staff of maybe a
- 26:42Speaker 1: dozen people, and then at certain busy times of the
- 26:46Speaker 1: year they'll bring in maybe a hundred laborers to come
- 26:48Speaker 1: in and help with the harvest. I think everybody would
- 26:52Speaker 1: be better off if that orchard had a year round
- 26:55Speaker 1: staff of twenty people that were productive all year long.
- 26:59Speaker 1: And we're able to use technology to even out these
- 27:01Speaker 1: bumps in the labor demand. And so those people, those
- 27:04Speaker 1: twenty people are going to need to be higher skilled,
- 27:06Speaker 1: but they're also going to get paid more, and they're
- 27:08Speaker 1: also going to have more comfortable jobs, and overall they
- 27:12Speaker 1: will produce more per person than they would in the
- 27:14Speaker 1: other system. Of course, the lingering question is what happens
- 27:19Speaker 1: to the eight people who no longer have a job,
- 27:21Speaker 1: and who gets to enjoy the fruits of this more
- 27:23Speaker 1: efficient system. Technology has improved lives all around the world
- 27:27Speaker 1: and lifted millions out of poverty, but it is also
- 27:31Speaker 1: dramatically enriched an extremely small number of people. We mentioned
- 27:35Speaker 1: Elon Musk's neural link earlier, and he's not alone in
- 27:38Speaker 1: the Silicon Valley elite investing in transhumanist technologies. That should
- 27:42Speaker 1: give us pause, remembering what Ian Bremer said about cognitive differentiation.
- 27:47Speaker 1: So there's much to fear, and there are no obvious
- 27:50Speaker 1: solutions in sight, and yet people like Kai Fu Lee
- 27:53Speaker 1: and Gil Pratt, people who are leading the field, remain optimistic.
- 27:58Speaker 1: I wanted to know why there is a strong belief
- 28:02Speaker 1: that thought leaders should do the best they can do
- 28:06Speaker 1: to project a possible future and strive towards it and
- 28:11Speaker 1: encourage other people to help make that a reality. Because
- 28:15Speaker 1: whether we point at the future that is an utopia
- 28:19Speaker 1: or dystopia, if everybody believes in it, then it becomes
- 28:24Speaker 1: a self fulfilling prophecy. So I'd like to be part
- 28:28Speaker 1: of that force which points towards more of utopian direction.
- 28:33Speaker 1: Even though I fully understand and recognize the possibility and
- 28:37Speaker 1: risks of the negative ending, we will want to believe
- 28:42Speaker 1: in that utopian direction, honesting automation to help feed the
- 28:45Speaker 1: world without stripping ourselves of community interaction, because man cannot
- 28:49Speaker 1: live on bread alone, and we need to make sure
- 28:52Speaker 1: to balance gains inefficiency with preserving the fabric of our society.
- 28:57Speaker 1: In the next episode, we travel from the farm yard
- 28:59Speaker 1: to the battlefield. We meet some of the people pioneering
- 29:02Speaker 1: the use of AI and robotics to wage different kinds
- 29:05Speaker 1: of wars, and we speak with Arti Pravaca, the former
- 29:08Speaker 1: head of Darper, the agency that created the Internet, about
- 29:11Speaker 1: how technology is revolutionizing the military. I'm oz veloshin see
- 29:15Speaker 1: you next time. Sleepwalkers is a production of I Heart
- 29:30Speaker 1: Radio and Unusual Productions for the latest AI news live
- 29:34Speaker 1: interviews and behind the scenes footage. Find us on Instagram,
- 29:38Speaker 1: at Sleepwalkers podcast or at Sleepwalker's podcast dot com. Sleepwalkers
- 29:44Speaker 1: is hosted by me Oz Veloshin and co hosted by
- 29:46Speaker 1: me Kara Price. We're produced by Julian Weller with help
- 29:49Speaker 1: from Jacopo Penzo and Taylor Chikoin. Mixing by Tristan McNeil
- 29:53Speaker 1: and Julian Weller. Our story editor is Matthew Riddle. Recording
- 29:56Speaker 1: assistance this episode from Walter Kowski. Sleep Workers is executive
- 30:00Speaker 1: produced by me Ozvaloshen and mangesh had Tigella. For more
- 30:04Speaker 1: podcasts from my Heart Radio, visit the i heart Radio app,
- 30:07Speaker 1: Apple Podcasts, or wherever you listen to your favorite shows.