← On with Kara Swisher
Loading summary
Transcript95 lines
- [00:02]
A
It's on.
- [00:13]
B
Hi everyone from New York Magazine and the Vox Media Podcast network. This is on with Kara Swisher. And I'm Kara Swisher. Today we've got a conversation between futurist Martin Ford and economist Betsy Stevenson on the effects of artificial intelligence on employment. Ford has been thinking and writing about the effects of AI on jobs long before it was in the Zeitgeist. He's the founder of a software development firm and the author of four books, including the New York Times best selling Rise of the Technology and the Threat of a Jobless Future, which won the 2015 Financial Times McKinsey Business Book of the Year Award. Stevenson is a labor economist who researches the effect of AI on work. She was the chief economist of the U.S. department of labor from 2010 to 2011 and then served on the Count Council of Economic Advisors from 2013 to 2015. She's a professor of Public Policy and Economics at the University of Michigan and serves on the Executive Committee of the American Economic Association. Martin and Betsy are two very deep thinkers on this topic and couldn't be more vital. A quick note before we get to the episode. I'm doing a live Pivot Tour with my co host, Scott Galloway. The tour runs November 8th through the 14th and it hits Toronto, Boston, New York, D.C. chicago, San Francisco, and LA. We are sold out in several of those cities and almost sold out most of them, but tickets are still available@pivottour.com in some cities. All right, this is a really sharp conversation, so stick around.
- [01:58]
C
Support for the show comes from smartsheet. In today's episode, we debate the future of work and what it looks like with AI. But while the experts debate, your team needs to get work done. Smartsheet is the enterprise grade platform that helps you rise above the uncertainty. It lets you harness the power of AI to thrive in a changing world, and it's designed to minimize disruption. Now you can adapt, collaborate, and work with flow, no matter what the future holds. Go to smartsheet.com box to learn more.
- [02:32]
A
It is over.
- [02:34]
C
Martin and Betsy, thanks for coming on on.
- [02:37]
A
It's great to talk with you today.
- [02:38]
D
Thanks for having me.
- [02:39]
B
Now, Martin, you're relatively pessimistic about how AI will affect employment. And Betsy, you're relatively optimistic. So let's start by having you both lay your cards on the table. Betsy, why do you see the AI.
- [02:51]
C
Glass as half full?
- [02:52]
B
And then, Martin, why do you see it as half empty, so to speak?
- [02:55]
A
Well, economic growth, true economic growth, comes from only one thing, technological progress and so if we want to be able to produce more using fewer resources, and those resources include human labor, energy, capital, the only way to do it is through technological progress. AI is obviously a huge leap forward in technological progress, but I don't want to overstate my optimism. In almost every time of huge technological progress, it has moved us to a better era, but it's caused a lot of pain along the way. I always tell my students, economists describe the Industrial Revolution as if it was fantastic. Obviously they don't read a lot of Charles Dickens. So there is definitely a downside. And I think managing the downside is the task for society and for policymakers.
- [03:54]
D
Okay, Martin, I basically agree with what Betsy said. It's just that my position has been, since I started writing about this 16 years ago, is that ultimately it's inevitable that machines are going to displace, I think, probably at least half the workforce, maybe more, because AI and robotics are going to get to the point where it matches and then exceeds the capability of most typical average workers. That doesn't mean that no one is going to have a job. There'll be some people that have particular talents and personality traits and things like that. But I think for the majority of people that come to work and do relatively routine, predictable things, we will reach the point where it's going to be very difficult for those people to find a market for their skills. I think the machines will simply be better at just about everything that they can do, and we're going to have to adapt to that.
- [04:46]
B
I'm glad you mentioned robotics, because often it's left out of it because this is the way that progress is going is also pretty significant.
- [04:54]
D
Absolutely. AI seems to be the focus right now, and maybe initially we worry more about white collar jobs, but absolutely, robotics is making enormous progress. So certainly in environments like Amazon Warehouses, I think it's going to be quite disruptive.
- [05:09]
B
I've been to those warehouses and many, even a decade ago, you could see which way it was headed in terms of getting rid of workers. But Walmart CEO Doug McMillan recently said that, quote, AI is going to change every job. Is that accurate or hyperbolic? Is AI a truly revolutionary technology or radically transformed how we work? Or is it more like the Internet, which has brought. It brought about massive changes, but it hasn't fundamentally changed how a lawyer, accountant.
- [05:36]
C
Or plumber does their job.
- [05:38]
B
Martin, let's hear from you first. And then Betsy.
- [05:40]
D
Yes, I think that AI is unique. I think it's unlike any other technology. And the reason is that it's going to directly displace what is our core competence, right? The thing that sets us apart is our intelligence, right? Our cognitive capability. That is what so far has always allowed us to adapt and remain relevant to the workforce, even in the face of advancing technology. But I think what's happening is that finally we're going to get a technology that comes directly at our comparative advantage. It's going to displace the thing that really sets us apart.
- [06:13]
A
Betsy so it's worth thinking about other big changes, because what we used to do for a living was agriculture. Almost all people on the planet worked in agriculture, and the share of the planet engaged in agriculture has absolutely plummeted. And that hasn't meant that we don't have things to do. It doesn't mean that we don't have work. It does mean we don't work in agriculture. So I think when I talk about the importance of the transition, it's because we know that there will be jobs that are better able to be done by machines. The question will be, what are the new things that arrive and what's the pace at which they arrive? And do we like those new things? Are they better jobs? In a lot of ways, the jobs people initially left the farms for in factories were not necessarily better at first, and they didn't really improve living standards. But I think most of us sitting around, right, you're doing a podcast for a living. I'm teaching at a university. I would much rather that than working in the fields. So our jobs have gotten a lot better. But transformation takes time. So I actually am gonna push back on your idea that the Internet didn't change fundamentally how I do my job as an economist or how a lawyer does their job. I mean, the number of papers academics write nowadays because of computing. So let's put computing in the big bucket. But also because of the Internet, because I can access. I can download the census during this podcast if I would like, when, you know, only a couple generations ago, if they wanted to get access to census data, it was a very big progress that involved punch cards. So it has fundamentally changed our productivity, and I think, in a way, made the job a lot better. But technology is not guaranteed to make the jobs better. And I think that's where we need to have our discussions and our focus about what are the tasks that are going to be replaced by machines and how do we feel about those tasks. But let me say one last thing. Before we get replacement, there's something else that's going to happen, which is that the wages are going to Change, because it's not enough for the machine to be able to do the job. The machine has to be able to do the job cheaper than a human. And one thing that could happen is humans could just do the job for less. And so what we could see happening is the wages of a lot of white collar workers facing downward pressure. And that, I think is problematic, but I think that that's much more likely to happen first before we see, you know, wide swaths of 50% of the population unemployed. What we're going to see is real reductions in the compensation people get. You know, there are still people using manual carpet weaving looms today because they're cheaper than using a mechanized one. So it is the interaction of the machine being able to do it and then being able, the machine doing it cheaper than any human will do it.
- [09:31]
B
Martin, will you answer that? The historical element of this happening is that it's just a shift in jobs, which is essentially what she's saying.
- [09:40]
D
Right. But I think that largely it's been a shift from one sector to another. So as Betsy said, we all used to work on farms, and people were working on farms and they were doing relatively routine, predictable things. Those jobs disappeared when agriculture mechanized and there was disruption, but there was a transition largely into manufacturing. Right. So people went from farms, they went to factories where they were doing relatively routine, predictable things. Then manufacturing, also automated and offshore. And now in countries like the United States, most people are in the service sector, but they're still doing relatively predictable, routine things. And the point is that this time we're not talking about a specific technology like agricultural machines. We're talking about a general purpose technology, and we're talking about a technology that is going to directly displace, you know, our cognitive ability. And what that means is going to be deployed everywhere. So it's hard for me to see, maybe it's a lack of imagination on my part, some entirely new sector arising at this point. You know, we saw a manufacturing rise after agriculture, then services, but AI will be deployed in all of those sectors.
- [10:53]
B
So there's nowhere to go.
- [10:54]
D
Is there going to be some entirely new part of the economy that can absorb tens of millions of workers that are going to be potentially displaced from those other jobs? I doubt it, because AI is going to be used everywhere.
- [11:06]
A
So I want to both agree with one point and disagree with another. We've seen general purpose technologies before. I call electricity general purpose. We use electricity for all sorts of things. It's not specific. I think the way to think about this is we have a technology that is going to replace our intellect. And that I think is what is so frightening for people. If you look back at that general purpose technology of like the steam engine, electricity, power, what did it replace? It replaced our brawn. For anybody who has kids, you might have seen the movie Beauty and the Beast. Gaston was the most popular guy in town because he was strong, right? Gaston ate all those eggs so he could lift all those carts and oxes. And that's because strength was valued before we had technology that could replace strength. Now technology is superhuman strength. Being the strongest guy is just not that different to being the weakest guy compared to being a forklift. And the problem with AI is right now there's a bunch of people who have set up their status in society and their income in society by being 140 or 150 IQ person versus being 100 IQ person. And that's going to be the same thing as being like the really strong guy versus the average strength guy in the future. It's just going to be swamped when we have machines that are all at 300 IQ points. And so what will people, what will our comparative advantage do? And I think none of us can really forecast the future in quite the right way to say where will we find the bonds of reciprocity between humans that will create a space both for the ways we spend our time, the ways we get status and the ways we earn an income.
- [13:12]
B
Right. So speaking of forecasting how we forecast the future and how it's impacted by innovation, the Nobel Committee just awarded the Nobel Prize in economics. No, not to Donald Trump, but to an economic historian and two economists whose research looks at creative destruction, or simplest term, how innovation replaces older technologies that can lead to economic growth. One of the winners, Philippe Aguion, actually co chaired France's Artificial Intelligence Commission and it published a report that says AI replaces tasks, not jobs, and that the spread of AI will create jobs.
- [13:47]
C
So is he right?
- [13:48]
B
Will AI end up being a job creator or destroyer? Why don't you start Martin and then Betsy?
- [13:55]
D
Obviously it will do both. It absolutely will create jobs and it will destroy jobs. The question is, will there be enough new jobs created to absorb people that lose jobs? And secondly, are those new jobs going to be appropriate for those people? Are they going to have the necessary skill set? I'm very concerned that relative to most people, you're going to over long run see more destruction and creation. And you know, it's true that it's that AI automates tasks, but I think that that's going to lead to organizations sort of redefining the boundaries between jobs. I think you can imagine two workers doing basically the same thing and half of what each of those people are doing gets automated in terms of the tasks. And then of course at some point management will look at that and say, you know what, there's really only one job there. Right. Because half of what's being doing can be automated. I think that's the process. It's rarely going to be a one to one correspondence between an algorithm or a robot and an existing job.
- [14:56]
B
Right. What do you think, Betsy?
- [14:59]
A
So here is the conundrum to think about. AI is going to allow us to produce more with less. That makes us richer, that makes us want to consume more. That means we need more of all the inputs, which includes humans. And, and that's where the sort of standard economic thinking goes is yes, AI is going to replace tasks, but by being a cheap way to produce, it's going to allow us all to consume more. And so we'll scale up and that's what's going to lead to there being enough jobs for humans. There's a lot of worry that maybe this time will be different. And I, I think I don't have a worry this time will be different. I do think you eventually get to that point. The question is, does this process happen so fast and do we fail to redistribute all those income gains in a way that means that demand goes up as well? Right. If Sam Altman is the only one who gets rich off of AI, then we don't actually need to produce more stuff because there's only so much Sam Altman can consume. That's sort of that vision of the manager only needs half as many workers because they're trying to produce the same amount of stuff and they're now doing it much more efficiently. So the distribution problem cannot be fully divorced from the issue of how will this allow us to scale up and how will it allow us to consume? But also keep in mind that if we don't solve the distribution problem, that's its own limit on AI adoption. There's no real incentive to adopt AI if there's no demand.
- [16:44]
B
So we're starting to see the effects of AI unemployment, especially on young college graduates. The unemployment rate for that demographic is averaging 4.59% this year versus 3.25 in 2019. It's noteworthy, it's not earth shattering. When do you expect to see significant job losses? For example, unemployment peaked around 20%, 5% during the Great Depression, Martin is predicting 50%. So, Martin, why don't you start and explain why you think 50% is your number?
- [17:12]
D
Well, you know, that could be far in the future. I mean, when I wrote my first book 16 years ago, I thought it might be 50 years before something like that happened. I think we're clearly on a trajectory toward that, and we're at a moment in time where, as you say, there is pretty clearly an impact on entry level jobs. And I don't think it takes that much to start building systems that will go beyond that. For one thing, I don't think that it requires artificial general intelligence or AGI or certainly super intelligence, which is what everyone's talking about now. And again, the reason is that most people are doing relatively predictable, routine things. They might be very smart people doing analytically difficult things, but not so difficult for AI. I think that, you know, given where LLMs are now, if they can address the issue of hallucinations, and they can also build what they're calling agents, right, that can plan and execute complex multitask operations and can also utilize the kind of tools that the white collar workers use, the software, the digital tools, I think that we could begin to see an impact quite soon, I would imagine. It's really a question of what happens with the technology and how fast they can address these issues.
- [18:30]
B
Betsy.
- [18:32]
A
So first, I just want to put it in perspective and say that even at 4.5%, the unemployment rate of recent college grads is lower than the unemployment rate of high school grads. College grads have always had an enormous advantage in having an extraordinarily low unemployment rate that is traditionally about a third of what the unemployment rate is for everybody else. So it's possible that college grads start to look like everybody else, but that's not what I think will happen. What I think is going to happen. And now I'll get a little wonky on you, which is I think we're going to change what the age wage profile looks like. So if you are a college grad, you tend to earn less in your early 20s than even a high school grad. Why? Because you're focusing on college, and the high school grad who's gone straight to work is focusing on work. But by Your sort of mid-20s, your wages are growing faster and you overtake those high school grads and they keep growing, and your wages don't really peak until around 40, around age 45, that it's not extremely sharp, but it's just in your 40s you earn more than in your 20s. I think that that steepness though, that difference between what you earn in your 20s and what you earn in your 30s or 40s is gonna get even bigger. And it's because what's going to happen is a lot of the things that that 22, 23 year old, recent college grad does is going to be able to be done by the more experienced person with their AI tools.
- [20:07]
B
You don't need a young lawyer, for example.
- [20:09]
A
And so then the question will be, what do we set up to help the young people get the experience they need so that they can be an experienced lawyer? And it may be that we need to have more training programs that what you're actually going into is a new type of internship where you know, you're really just learning and you're getting a stipend that maybe you pay back out of future earnings, or maybe the government sponsors these stipends. Think about what we do with medical residents. You go in as an intern in the hospital, you're not that productive. In fact, you're subtracting because the attending is teaching you. And so we get government sponsoring those kinds of first positions for medical students in medical settings. And we may need that for lawyers for all sorts of white collar jobs where it's just really hard to be super productive when you're young. Now the thing that fights that is the faster technology changes, I can tell you the young people are better with it. They pick up very quickly, they move very quickly. But I think we do need to worry about generational gaps because I think that AI is going to make it very easy for sort of that Gen X, very young boomers to hang on for quite some time, being very productive and using the experience they have.
- [21:37]
B
Yeah, because they don't need those associates say if it was a law firm, they don't need what they're doing, which is mostly easily replicated by AI. For example, I'm going to be doing a live taping of on with Kara Swisher at Smartsheets Engage Conference, which runs from November 5th through the 7th in Seattle. You can register for their conference on their website and you'll hear more about smartsheet services coming up in the break.
- [22:11]
C
Support for this show comes from smartsheet. The conversation around AI is ever changing and complicated. Even in the last five years, what's possible with AI has changed drastically. AI has been around a long time. People don't realize that AI has always been in your Internet and it's been.
- [22:28]
B
Used to give you all kinds of.
- [22:29]
C
Things you didn't realize. But what's changed is the intelligence has moved in ways that are like a Cambrian explosion. And now it's entered the scene as a consumer product. Now the thing is, they think AI will solve every problem we'll have in the coming years. But beyond the talk, there's also what's possible today. Smartsheet is using AI in a way that's strategic, efficient and empowering for businesses. Smartsheet has a suite of AI capabilities which includes intelligent assistants and expert AI agents. It's designed to help you automatically the time consuming and mundane tasks and focus on big ideas. Smartsheet AI provides proactive guidance, flags risks before they become blockers, and helps you make data driven decisions with confidence and control. It's about making every job more impactful and every day more fulfilling. Go to smartsheet.com Vox to learn more.
- [23:26]
B
So, as we mentioned earlier, according to conventional wisdom, white collar jobs are more affected by AI than blue collar jobs, which implies the jobs people do with their hands are relatively safe from AI.
- [23:36]
C
Is that accurate?
- [23:37]
B
And if so, will that always be the case? You mentioned robotics at the beginning, Martin, because I think it gets a short shrift in terms of its development. These advanced robots and they don't have to necessarily look like humans, actually, those are more complex robots. But you know, like a dog walking robot doesn't have to look like a person, it can look like a wagon. As anyone who's gone inside of Amazon Warehouse sees, they're going to be able to replicate everything. When I saw one of the things they were doing a couple years ago, I thought, oh, they don't. You're not going to need any of these people. Soon you could see where it was headed. So Martin and then Betsy, because you had brought up that idea of robotics pairing with AI.
- [24:17]
A
Yeah.
- [24:18]
D
I think that robotics are going to have a huge impact, especially in environments that are very controlled and predictable. And Amazon Warehouse is a good example of that.
- [24:28]
A
Right.
- [24:28]
D
It's walled in. Amazon controls exactly what happens in its warehouse. It can separate the people from the robots so you don't have collisions. It can reorganize the flow of packages and products through the facility. Right. That's completely different than the outside world where you got to deal with chaos. And that's why, of course, you know, self driving cars have taken a lot longer than many people expected. But within those controlled environments, warehouses, factories, I think eventually, maybe even supermarkets, fast food, restaurants, I think the robots are going to have a dramatic impact. So it's Definitely not just white collar jobs that are going to be impacted. It's going to be a very broad based concern.
- [25:11]
B
I think Bessie, talk about this because one of the things I remember from visiting this Amazon warehouse, these robots were putting all kinds of different things in boxes and then moving them around via this system of logistics. So you'd have a book by Sheryl Sandberg next to Hairspray, next to combs, next to, you know, a fleece jacket in one tub. And I was, I asked why are they doing that? And they said they know where these things are going to go. And they weren't individual orders, they were larger. It was really interesting and this was early on, so it was more efficient because it thought differently than a person would. Right, so talk a little bit about this.
- [25:51]
A
Oh well, I think that absolutely those jobs are going to get mechanized away. And I think that's a job where, you know, I don't want to dismiss anyone's work. Everyone's, you know, labor is valuable. But I think those are the kinds of jobs where, you know, there's other places where we need people in society. And I, my hope is that the increased wealth that we get from AI will mean we can pay caregivers more. But I actually want to give you a concrete example around self driving cars. So self driving cars are coming, but let's think about how they operate. Right now they operate best in very dense cities. Why? Because they're expensive capital and therefore they work best when they can be deployed and used for a large number of hours in the day. Now the problem is there aren't as many people wanting to go places at 3am as there are people wanting to go somewhere at 6am or at 5pm and so if you have enough self driving cars to get everybody where they need to go, even in a highly dense area, you're going to have a bunch of those self driving cars sitting around doing nothing for most of the day. And so I think it'll be a very long time, even in a dense city before we end up losing the human drivers because of the fact that humans are going to be much more flexible and willing to step in during the peak times. So that's one way in which you can think about capital working with humans. Now you might say, but the human driver has a car sitting in their garage. That's true. So as long as people have a preference to own their own car and we don't all turn to self driving cars, that will be the scenario that we're in for quite a long Time and it'll take even longer for these self driving cars to get out to less dense areas because the cars will not spend enough time, you know, with, with people in it.
- [27:50]
B
That's a fair point. So let's pivot and talk about how government should help citizens deal with job losses or think of, you know, how to make that transition. Should the federal governments take steps to regulate the use of IA specifically to prevent job losses or at least slow them down? Or would that inevitably just help, say, China beat the US in AI? What is the government's role here? Martin? And then Betsy.
- [28:14]
D
No, I definitely don't think we should have laws preventing automation. We do see that now, right? The unions at the ports are opposed to automation and so forth. And that just holds us back, that holds back progress and it certainly does make us less competitive with other countries. I think we're going to have to figure out this inevitable transition. You know, trying to just slow it down is not the answer. We've got to figure out how to adapt to it. Certainly that can be in part investing in education, you know, community colleges, things like that, so that people can pick up new skills and to the extent they're, they're capable of it, you know, transition into new positions. But in the long run, I think that ultimately we're going to have to go beyond that and we really need a new paradigm. And that's why I've written a lot about, for example, universal basic income, a guaranteed minimum income or negative income tax, those kinds of approaches where we're ultimately going to need the government to directly supplement incomes, not just so that people can survive, but so they can be consumers. Right? They can go out and create demand in the economy so we can continue to have a vibrant economy. Or as Betsy suggested, you know, if Jeff Bezos is getting all the money, you know, he can't drive the whole economy just by buying another super yacht. Right?
- [29:37]
B
Well, he looks like he's doing a good job right now, but Betsy.
- [29:42]
A
So Martin, I'm gonna guess that I can get you to agree with this, even though it is about slowing things down. I think the first thing we have to do is dramatically reduce taxation on labor and move our tax system either towards consumption, like goods, or towards capital. So historically the view has been we need capital investment to have economic growth. And capital is much more flighty. If you tax it, people will just stop investing and they'll consume instead and people will just keep working no matter how much you tax them. So we'll tax the people, the labor, but we won't tax the equipment very much. In fact, we might even subsidize the capital investments. And it's even true that we subsidize, you know, investment in technology, but we don't subsidize investment in workers like human capital, worker training programs. There's been bills on the Hill that have tried to change that that have just gotten no traction or not enough traction. And I think the smart way to slow it down would be, wait a minute, why are we trying to discourage human labor at a time where we think human labor might be being replaced? And why are we trying to speed up that replace by having a tax system that encourages you as a company to move away from people and towards capital? And so that's like a first step would be to change our tax system to prioritize people keeping jobs over over machines.
- [31:22]
B
Bernie Sanders is calling for a so called robot tax on corporations that use AI to eliminate jobs in order to raise money to help people affected by that technology. What do you each think of this approach? Martin first and then Betsy.
- [31:35]
D
I'm not sure a specific robot tax makes sense because it's going to be very hard to determine exactly what is a robot, what is AI that's displacing jobs? I think, as Betsy said, robots and artificial intelligence are capital. So probably it makes sense to more generally shift our taxation toward capital. But of course, all of this is a staggering political challenge. I mean, basically all the ideas we're talking about here seem extraordinarily difficult to ever make it happen.
- [32:05]
B
Betsy.
- [32:06]
A
I agree with Martin that what's a robot? Is a challenging question. So there's an implementation question there. And I think ultimately you pass a robot tax and everything will start to count as a robot. So maybe that's just the way politically you get it done. I think that we have to figure out how we are going to switch priorities or at least put workers and machines on, you know, a level playing field. And for sure something's got to give there.
- [32:39]
B
We'll be back in a minute.
- [32:50]
C
Support for this show comes from smartsheet. There are a lot of misconceptions about AI, that it's going to kill us or maybe just solve all our problems. Neither of those things are true. And that's probably the biggest misunderstanding about AI and how it can help our world. Like a lot of things, it can be both a tool or it can be a weapon. And in the wrong hands, it could be disastrous. In the right hands, it could change everything from drug discovery to cancer research to all Manner of ways we work to help us be smarter at work. Smartsheet wants to help you move past the confusion and get to work. They offer an AI that's strategic, efficient and empowering to business owners. Its suite of AI capabilities includes intelligent assistants and expert AI agents. These can take on the mundane tasks that slow you down and let you focus on the bigger picture. And smartsheet AI provides proactive guidance, flags risks before they become blockers, and helps you make data driven decisions with confidence and control. It's about making every job more impactful and every day more fulfilling. Go to smartsheet.com Vox to learn more.
- [34:01]
B
So Martin, you talked about UBI necessary to replace lost income from job losses caused by AI. And Betsy, you've suggested that a digital dividend might be better. Could you explain what each of those things that are and how they're actually different? Betsy and then Martin so when I.
- [34:16]
A
Say digital dividend, I think as an economist I was often put off by the idea of a universal basic income. It's sort of like I've given up on you. We're just going to give you an income so you can sort of bugger off and leave us alone. And what I've been trying to think a lot about is that the way the world works right now is we're all born with 24 hours in the day, a brain and physical strength. And our physical strength is worth something, but not as much as it used to be. And our brain is now worth a lot, but it's gonna be worth less in the future. So if our brain's not worth that much and our physical strength's not worth that much, what do we have? And what I don't want is a world in which what kids basically have is whatever their parents had. I'm really worried about us getting stuck where today's distribution of income sets up the future distribution of income for the next hundred years. And so what do we all have? We all have our data. And I know the ship has kind of sailed on this, but I really think that government needs to stop and realize that our data is the thing we have. It is something that companies value, they want it. And we should tax the use of data and give that money back to people. So it's going to sound a lot like a universal basic income. But instead of being like a handout, now what I'm talking about is something that's based on you're contributing your data into the system and that you get a dividend from that. Another way to think about it is we could just nationalize some of our artificial intelligence and then pay a dividend on that. I think it's that the American public deserves a share, something that they have contributed to through their own data. Right.
- [36:09]
B
You know, one of the things that I had talked about for years was the idea of being paid for your data. I was sitting in a room with Steve Case. He was saying that he made $50 from every customer, essentially using their system and inputting their data. And I put up my hand, I said, where's my $25? Like, why don't I get part? You know, essentially, I'm your Soylent.
- [36:30]
A
Yeah.
- [36:31]
B
And so they don't think like that. These tech people think your data is their data, it's theirs for good. And in fact, so much so that even Sam Altman's saying, you can opt out of our system rather than opt in or be paid for that.
- [36:44]
A
That's because they know people won't opt out. And it is. It's really greedy. I hate to throw this around, but it's a form of slavery. Right. Like, it's no different than them saying, I have a right to your labor. They don't have a right to my labor.
- [36:58]
B
Right, that's correct. They don't have a right to most things, but they take it and then hope that people won't sue. But they will, in this case. Martin, talk about ubi, how that compares, how Betsy's ideas are different.
- [37:10]
D
Well, I. I'm certainly open to what Betsy is suggesting, and I think that if you're going to offer people a. A data dividend that is the same for everyone, then essentially that's the same as the ubi. Maybe it's a better branding. Another branding is a citizen's dividend. Right. Which is, I think, what Andrew Yang had when he ran for president. Or is a freedom dividend, something like that. So it's just branding. That's fine. However, if you're going to have a scheme, and I've heard this sometimes too, where the amount of money that you would get is actually somehow tied to the value of the data that you're generating, then I think it would probably set up all kinds of perverse incentives. Right. People would try to game it, try to actually go out and generate data in some way that results in a higher income, and that would essentially destroy the value of the data. Right. Because data is only valuable if it actually reflects your true preferences or behavior. So that would be my only concern. But if, you know, as far as if UBI is Bad branding, then for sure, call it something else.
- [38:17]
B
But no matter how you slice it, as Betsy pointed out, the world in which a significant chunk of labor is done by a means, the companies will become incredibly powerful and their owners even more than they are today. And so if the future belongs to whoever controls AI, and most of the people who control AI have publicly aligned themselves with Donald Trump, then how do we make sure that citizens still have a say in how our society functions if they can't go to dinner on the Mar A Lago patio at the White House?
- [38:45]
D
Well, I think that it will be more democratized than we think. I mean, it's true that there are a few companies that are actually building these systems, but the systems are going to be available to everyone. Right. They're integrated into cloud computing. And I think that what you'll see in time is that speaking in terms of businesses, the value from AI is going to be distributed across industries. It's going to be the people that leverage it, that capture most of the value, not the people that are necessarily building it. Because you see, at least now, three major competitors. Right. There's, there's certainly going to be intensive competition there.
- [39:23]
B
Well, except they're all doing round tripping deals with each other to move $5 back and forth, Right?
- [39:28]
D
Exactly. And if we end up with a burst bubble, that itself will create more unemployment than direct automation, for sure, Betsy.
- [39:39]
A
So keep in mind that as a resource becomes plentiful, it becomes cheaper. And if AI is everywhere, it will be the cheap thing. And if you need a human in any place, humans will become the place of scarcity and we could end up seeing the problem get solved that way. I do want to say, though, I am a huge believer in markets and capitalism, but in capitalism in its pure form, like very competitive markets. But you have to keep in mind that AI is a little bit different. The spillovers are everywhere. It literally couldn't train it without having taken data from other people, without turning to the public commons. Right. So they are relying on a lot of public help, and that makes them different from your butcher, your baker, and that sort of standard competitive framework. The thing that worries me most about what's happening in the United States is we're moving even further away from a competitive market and we're moving towards a market where the people who curry favor with the politicians get rewarded. And that does not work very well when you're in even a world where it's butchers and bakers. It's going to work even worse in a world with AI I don't think that's good for anybody. I think that that will limit our potential. I think it could cause the whole thing to sort of implode. And, you know, the countries that have robust democratic institutions are going to get through this much, much better. And right now, as the United States has every one of its institutions under threat, I think we have a lot at risk right now. That is the interaction of a collapsing democratic system with a rising power in AI. And that's going to be a challenge for the United States.
- [41:51]
B
Yeah, I think the idea that small companies can't bubble up because they're not at the table like pigs at the trough is a real problem. I've used the term pigs at the trough a lot, but that's what it feels like in a lot of ways. So how should people find meaning in their ability to provide for themselves if they it's no longer tied to work. Betsy, then Martin.
- [42:12]
A
You know, there is a concept in Japan called ikigai. An ikigai is your meaning or purpose. And if you look at the societies where Japanese citizens live to extremely long ages, what you find is everybody has an ikigai, but their ikigai does not usually work. And in fact, Japan recently did a survey of the entire population and, and what they found was life satisfaction is very correlated with your work or your income, but your ikigai is much less correlated with your income or your work. And in fact, what ikigai is most correlated with is your participation in civic society and volunteer groups in community based things. So people who are involved in four or five organizations have the most ikigai. And I'm, you know, I'm the mom of teenagers and I have spent a lot of weekends at dance competitions with my oldest daughter. You know, dance competitions are not because we're dealing with the most elite dancers. We're dealing with high school girls and boys who really like to dance, who like to compete, who have a lot of fun and they put their all into it. And there's a lot of ikigai that comes from that. So I think there is a lot of space for us. But I want to give another warning about the United States compared to any other country, which is we have seen our happiness and our life satisfaction stagnate over the last 50 years. Why? Because we're no longer doing as many things together and civic organizations. Bob Putnam has this book that's 30 years old now, Bowling Alone, that we still bowl, but we don't do it in a league. We're not getting together with, you know, broad parts of the community. And we're not using our wealth to build a safety net, to add vacation, to do all sorts of things that can make us happier. And so I worry a lot about whether the US has the capacity to respond to this in a way where we're all going to be able to find our ikigai. But it's certainly possible for people to.
- [44:29]
B
Martin?
- [44:30]
D
Yeah, it's definitely a huge problem because jobs don't just provide income, right? They provide a sense of dignity, a sense of self worth. They occupy time, they give people a sense of meaning. And we're going to have to find replacements for all of those things in a world where people aren't working as much. One idea that I have put forth is that, that maybe instead of just having one level ubi, one level payment, that everyone gets the same thing, there can actually be the opportunity to earn more money or to receive more money if you do certain things right. If you, for example, stay in school and study, if you go and work in the community, maybe it should be possible to actually get a higher income. I worry especially about education. I mean, imagine that you're a high school student that's maybe struggling a bit in school, you're in danger of dropping out, and you know that when you turn 18 or 21, whatever, you're going to get exactly the same universal basic income as everyone else, regardless of whether you stay in school. Okay. I think that that's really a perverse incentive. So I would actually say to people a higher income if they remain in school. So I think there are ideas to work with there. But again, it's going to be a big challenge in many dimensions.
- [45:48]
A
I was just going to give you one really interesting study which showed that for people in manual labor jobs, their sense of meaning and purpose in the world went up when they retired. And that's because they were able to shift away from work that didn't provide a lot of meaning to things that did provide meaning. And so I know, particularly for modern knowledge workers in the United States, the idea that meaning and dignity is associated with paid labor is very, you know, is a common thought. But I'd like to remind you that lots of women did not work in the paid labor force and they had plenty of dignity because they had a role in society. So it is important that we remember that you have to have a role for sure, but it doesn't have to be paid work. And it historically has not always been paid work for all people. Right?
- [46:42]
B
That's absolutely True. Fair point. So let's end by talking about the role of formal education in an AI driven society. Sam Altman recently said he is envious of Gen Z college dropouts because they have mental space to build startups. And then there's the growing skepticism that college degrees are worth the time and money that's propagated by Peter Thiel. As AI becomes more advanced, what happens to in person university degrees? Will they remain the gold standard? Martin, you were suggesting we pay people to stay in school. I've heard more and more young people say what's the point of college? So Martin and then Betsy, I hope.
- [47:18]
D
That college continues to be important to people. I think there's going to be an explosion of other alternatives. We're going to have online degrees that involve AI tutoring people and so forth. And some of those may rise to the point where they're a reasonable alternative to going to a four year college. But I also kind of think that especially as we look to a future where maybe there aren't going to be so many jobs, the intensely vocational bent of higher education is going to maybe shift back toward more like a general liberal arts type education. It just makes us better people, better citizens, more broadly and flexibly educated, rather than trying to focus on getting a job in a specific area. So that could actually be a positive thing if that happens. And as long as we can keep people motivated to continue to be interested in education, which is certainly going to be a challenge.
- [48:16]
A
Betsy, I love that so much because I really hate the vocational. What kind of job am I going to get from this education? Because education is supposed to teach you how to think, how to be not just a good citizen, but also how to learn so that you're a really good, active, lifelong learner. I love a liberal arts education, so I'm going to agree with Martin on that. I do think there's an issue of how do we keep people feeling motivated? Like what are they doing in school? What are they trying to accomplish? What's their goal? And I think there's a lot of unanswered questions there. And education itself is going through a pretty profound shift because of the amount of AI use in doing assignments means that education is becoming kind of meaningless when you're cheating through AI. And so everybody's sort of trying to figure out how they're going to adapt to that. So I think there's a lot of adapting to be done both by the universities and by society in terms of how do we think about education.
- [49:26]
B
If you were giving advice then to high school students about how to plan their careers. What would you tell them? Should they go to college? And if so, what majors do you think will be, I guess, least impacted by AI? I know you, you hate focusing on vocation, as you said, but it's important.
- [49:40]
A
You know, I think understanding people is going to be one of the most valuable skills out there. So you can never go wrong with psychology. I think that also means you can never really go wrong with economics, because economics gives you an analytic framework for thinking about how people make choices. I agree with Martin that this idea of a broad education is what you're trying to do. And I really push creativity. It's a great time to do musical theater, to think about art. In every society, when technology has come along to make us richer, we found more money to pay for art. And so, you know, let's build everybody's creative side out, because I think there's going to be a lot of space.
- [50:25]
B
For creativity just to even to amuse the rich people. Martin.
- [50:29]
D
Yeah, no, I agree with that. I, I don't think it matters that much specifically what you major in. What matters is the kind of skills that you pick up. And I think Betsy is absolutely right that you want to emphasize working with people and, and team building and, and this type of thing, because that, of course, is what AI is not good at, at least not yet. It's not good at it as well as genuine creativity. And it's very hard to predict anyway. I mean, you know, if you're a freshman and you decide to major in one thing four or five years from now, it could be a completely different scenarios. So I think that really focusing on those core skills and talents is what's going to be important. One other thing I'd add is that not everyone, of course, will go to college. And certainly, absolutely one of the best things to do now in the face of all of this is to enter one of the skilled trades. Electricians, plumbers. These jobs are really hard to automate. You would need C3PO from Star Wars. Right. A very advanced robot in order to automate those jobs. So. So that's quite a ways off.
- [51:36]
B
Yeah. It seems like AI is going to force us to rethink what it also means to be a human. What do you think we'll say about how we handled the rise of AI? What worries you the most and what makes you the most hopeful? Martin and then Betsy?
- [51:50]
D
Yeah, the thing I worry the most about is the economic impact. I'm not so much concerned with the doomer Scenario where it's actually going to wipe us out existentially, maybe some data, it'll be a real concern. But I think we're going to have massive economic and social disruption long before that. And it could come in the form of automation of jobs as we're talking about here. I think that's inevitable. But we could also run into a problem with the AI bubble. Right. Bursting, and I think that could be really devastating to the economy. So those are the two big fears I have over the next ten years or so. But the optimistic side of this, and I think Betsy and I would agree, is that AI will be the most powerful tool we've ever had to amplify our intelligence and our creativity. And that's going to allow us, hopefully, to make enormous progress across the board in all kinds of scientific fields in medicine, in technology, and that should, in the long run, make us all better off. But we got to figure out how to adapt to the dangers that come with this. Especially, as we said earlier, the distributional issue. How do we make sure that the benefits of this end up distributed to everyone and not just a few people?
- [53:08]
B
Betsy, get the last word?
- [53:10]
A
So I think there's a lot of agreement there. I'm really optimistic about a lot of the medical breakthroughs we could get, the living standard improvements we could get. What I'm actually most concerned about, though, is what's going on socially and how that starts to interact with AI and how it's already interacting with AI. You know, we already have vaccines that cure the measles, and we have people afraid to get them. So we can have all the cures we want, but if we have a society where we don't trust each other, we don't trust the government, we don't trust the medical profession, we're all going to end up in silos of aggression towards each other, and that could erode all of the benefits. And that actually worries me a lot more than the job loss. I do think there's a fear of some people getting left behind in the next 10 years. But my bigger fear is the loss of social cohesion, because that's already underway.
- [54:09]
B
Yeah, absolutely. We'll have all the good ideas, but we won't use them. We won't avail ourselves to them. I really appreciate both of you. It's really. It's a really interesting discussion, and we will get back to you in 10 years.
- [54:22]
A
I will look forward to my agent talking with your AI agent in 10 years.
- [54:28]
B
Okay. All right. Thank you, Martin, also.
- [54:31]
D
Thank you.
- [54:31]
A
Thank you.
- [54:36]
B
Today's show was produced by Christian Castor, Roselle, Kateri Oakum, Michelle Eloy, Megan Burney, Kaitlyn lynch and Devin Schwartz. Nishat Kurwa is Vox Media's Executive producer of podcasts. Special thanks to Annika Robbins. Our engineers are Fernando Aruda and Requond and our theme music is by Trackademics. If you're already following the show, start making space for more creativity. If not, you're just another pig at the trough. Go wherever you listen to podcasts, search for on with Kara Swisher and hit follow. Thanks for listening to on with Kara Swisher from Podium Media, New York Magazine, the Vox Media Podcast Network, and us.
- [55:12]
C
We'll be back on Monday with more. Thank you to Smartsheet for supporting this episode. What's the future of work? One thing is clear, adaptability is key. Smartsheet provides the tools to help you and your team not just survive, but thrive as things change. The enterprise grade AI powered platform of smartsheet empowers you to get more done and make a bigger impact. Find out more@smartsheet.com Vox that's smartsheet.com Vox.