Loading summary
Transcript33 lines
- [00:00]
A
This is an iHeart podcast. Guaranteed Human.
- [00:09]
B
On this episode of Newtsworld on February 9, Matt Schumer published an essay on X entitled Something Big Is Happening. In it, he outlined a future where AI would take over most computer based white collar jobs and encourage those in the workforce to start using the latest AI tools. Now, the essay was significant in that it went viral with over 80 million views and it raises the whole question about the emergence of artificial intelligence and the way in which that will affect the entire employment system. My guest today authored a piece entitled AI Transaction Costs and A Quiet Shift towards Self Employment, in which she details that quote, what gets far less attention is whether AI changes how work is organized in the first place. Here to discuss AI's impact on jobs, I'm really pleased to welcome my guest, Leah Palagashvili. She is a senior Research Fellow and Director of the Labor Policy Project at the Mercatus center and George Mason University. Her research focuses on labor regulations, the gig economy, and the changing nature of work. Cleo, welcome and thank you for joining me on KnitsWorld.
- [01:46]
A
Thank you, Speaker Gingrich, for having me.
- [01:49]
B
I have to ask you, what led you to study this?
- [01:52]
A
Well, I took an economics class in high school and it all started in high school and I fell in love with economics. So from then on forward, I studied economics, got really interested in labor markets and started studying labor regulations. Got my PhD in economics. Used to be a professor of economics actually. So I think economics really helps you understand the world and helps you understand labor market markets in particular in ways that most people don't really look at. And so I think it's an, and it's an incredible valuable tool.
- [02:25]
B
There are a lot of people who feel that artificial intelligence is something which is happening to them, that it's almost inevitable they can't do anything about it. There's a real interesting question here about how can ordinary people understand the scale of change we're going through? And can they in fact influence either the direction the technology takes or the direction that they take in responding to the technology?
- [02:53]
A
Yeah, so there's a lot going on with AI from a purely research perspective. We see that AI is reorganizing tasks, firms are using it, workers are using it, and the tasks that workers are doing at their jobs is changing. And I think AI can actually be really empowering for workers because instead of having to do a lot of mundane tasks that you had to do before, I have a research assistant who had to do a lot of coding and other kind of mundane tasks. She was able to give those tasks to some AI tools and then focus on things that she actually wanted to build on. So she wanted to get better at communication and writing and speaking in conferences and public appearances, and she was able to focus on that. So I think it's not just AI is happening to us, it's also about how are we using the tools to change the work that we're doing and is it making us more valuable, more productive, and allowing us to do the work that we want to do. As my research assistant was like, I don't want to do this data work anymore. I give that to AI, But I focus on the things that I really want to do, which is speaking, communicating, writing and so forth. And I think another thing that we're missing, and this gets to the post that I wrote on, is again, there's this fixation on AI and work debates about whether AI is replacing workers in existing jobs. But I think there's a deeper economic question about does AI change how work is organized in the first place? I think about an example of a consultant. So 10 years ago producing a client ready report might have required a whole company's internal research staff, editing support, data visualization teams and so forth. And today there are AI tools that can help with drafting, coding, summarizing data, even formatting presentations. And that lowers what's called the minimum scale needed to operate independently. And what that means is that consultant can now become self employed and be an independent consultant and doesn't have to work inside of the firm. And the job isn't gone in this example, but the organizational structure around it can change. And I think that's an important aspect that we're missing, that AI can actually lead to more work being self employed and not necessarily inside of a firm structure as traditional employees do.
- [05:14]
B
You expect that people will in fact adjust and redefine the nature of work and redefine the way they organize work in response to the gradual evolution of artificial intelligence.
- [05:31]
A
That's a really great question. I think what might be happening and I can't predict things into the future very well, but I should say economists are bad at predicting things and recessions and so forth. But I want to give you a really clear example of what I see happening. AI is creating what's called a transaction cost. Technology shock and transaction costs are the real world frictions of using markets. So that means finding the right person, negotiating terms, verifying quality, ensuring you get what you paid for and firms exist, partly because it's sometimes cheaper to hire and manage people. Internally than to keep contracting everything out. However, with an AI shock and technology shock, it makes it a lot cheaper and easier to just contract out work and so on those terms, I think it is going to change how work is defined and how it is organized. And just to give you one clear example of how powerful this is, if you think about ride services, 20 years ago, if you wanted to find a driver outside of a traditional kind of limo or taxi company, you might have used Craigslist. But the transaction costs were huge, right? Because you didn't know who the driver was, you couldn't track them, payments were clunky and there was no reputation system. So as a result, not many people used Craigslist to find drivers. But then you had Uber come in and what Uber did, they didn't just create an app, it combined gps, digital payments, two sided rating systems to dramatically lower those coordination and trust costs. And suddenly it was easy and seamless to match someone who wanted a ride with someone willing to provide one. And that drop in transaction costs is what enabled the expansion of the gig economy over the past decade. Not because people suddenly changed preferences, but because technology made coordination much easier. And I think we're going to see the same thing with AI and transaction costs or a similar, but applied more to knowledge work, where we're going to see it be a lot easier to contract out, to do project on project basis. And for a consultant, a researcher who was previously in the firm, they now are able to do that same job outside of the firm with all of these AI tools. So you don't need the entire internal structure.
- [07:53]
B
There's this general fear that technology kills jobs. But the fact is, if you go back to 1800, 97% of the workforce was in agriculture. As recently as 1900, 41% of the workforce was in agriculture. Today it's around 2%. Yet we produce more food, greater range of choices. And the people who stayed in farming overall have done pretty well. But the people who love farming have also done pretty well, but at radically different things. Do you think that we're likely to see a similar pattern with artificial intelligence?
- [08:36]
A
Yes. Every tech shock in the past on Net has created more jobs. Now it's destroyed some jobs in some areas, but it's created new markets and new occupations that never existed before. So if you think about in the 1900s, there were not many airplane pilots, right now that's a whole industry in occupation that never existed over 100 years ago. And now it's an entire industry. And there are all of these examples of Jobs that existed in the 1900s that don't exist today. Now those jobs are destroyed by technological changes. But that didn't mean on net, that jobs are lost. On net. Actually, jobs grew and the economy expanded and markets grew and new industries emerged. And I believe based on the evidence that we've seen so far from past technological changes and from research on AI today, that's going to continue to happen. We're going to have new jobs and markets created that we can't even imagine what those are right now. Because if you ask someone in the 1900s what new jobs will be there, you know, in the year 2026, not many people would have responded. A bunch of airplane pilots, right, And a bunch of companies working on AI and workers doing those jobs in AI. So I think we really have to be a little bit more humble about, like, oh, you know, what are the new jobs? Well, we don't know, because every tech shot creates new markets, new industries and new opportunities. And on net, we're going to see more job growth, in my opinion.
- [10:10]
B
I actually know one of the guys who helped create automatic teller machines. And there was a study that said that ATMs didn't eliminate tellers. In fact, employment rose for tellers from 1980 to 2010 because as the operating cost for the branches went down, there were more local branches requiring more tellers, even if the branch had an ATM outside. It's just interesting how it doesn't quite work the way you expect it to.
- [10:42]
A
That's exactly right. We actually highlighted that ATM study on our substack as well in a different post about AI, because that is a great example, because what happened there is that more workers started to go into the relational roles at the banks instead of at the atm. And so I think that's a great example of how AI can actually augment workers and change the tasks and jobs that are done, not necessarily replace workers. And I think that's the key thing, but we need to understand with AI and labor markets is that the jobs, we can't say which jobs for sure will stay or which tasks will stay. But what we can say is likely the jobs and the tasks will change and there will be more jobs for workers. We just don't know what it'll look like and where it'll be and what those workers will do.
- [11:48]
B
As you look at these studies, I was kind of surprised to learn that in fact, the people who seem to gain the most productivity are the least skilled workers, while the best workers, the top performing workers, had the smallest increase in Productivity AI for some reason really helps train and empower the least skilled workers. But the people at the very top already have those skills, so they don't gain as big an advantage as actually the lower skilled workers.
- [12:19]
A
That's a great point. The evidence does show that when AI is used to augment human workers, helping them be more productive rather than outright replacing them, it can actually revive some middle class jobs and make workers more valuable in the labor market. So these lower skilled or less experienced workers often see, as you highlighted, huge productivity gains when I can support them. And I think that comes to a question that we've been hearing before too in these discussions about what will happen to middle class jobs or what will happen to the middle class. But Speaker Gingrich, the example you just used is a great one to highlight as actually it might help revive middle skilled jobs and make workers more valuable in the labor market. It's a really fascinating change and I should say that a lot of the higher skilled workers that I know are a little bit more concerned about AI versus lower skilled workers are like, this is great, like I'm being much more productive, I'm learning so much. So I really think it's an interesting dynamic that's happening here and I think it could speak a lot to this question of what will happen to middle skilled jobs. Because in the past we've seen some technological changes kind of hollow out some of those jobs. But I think AI has the potential to revive those middle skilled jobs, especially because of what we just talked about, that lower skilled or less experienced workers often see the biggest productivity gains. And this research by the way, is really fascinating. It's done by economists from Stanford and it uses some of the gold standard research methods. So it really showcases that we do see huge productivity gains, especially from lower skilled or at least experienced workers.
- [13:55]
B
This notion of the lower skilled workers, if you think about it, a lot of schools will say, you know, if you had 70% right, you get a C. Well, if you go to McDonald's and on the very first day you're the cashier and you get a 70% right, you get fired. You know, McDonald's actually wants you to be 100% right from day one. And so the standard of the real world, as opposed to what we've tolerated, is really different. And I've been intrigued. We've visited now several factories where you have people who come in, some as early as 14, learn how to be apprentices, get a job. It's seen as an investment in job training. We have met the CEO of a billion dollar corporation who entered the corporation as an apprentice, but they came all the way up. And my guess is that if you integrate it artificial intelligence into that kind of a system, you would be sort of dramatically empowering the starting apprentice and they would learn very fast and rise very quickly. But it's a very different way of really reshaping the relationship between information and actual work and being part of a group. The other part, which you've written about, is that we're going to lower dramatically the cost of being independent. Your artificial intelligence partner is going to be able to do so many things that today you need to be part of a team to get that done. You end up with a large number of people being independent entrepreneurs. Isn't that going to require rethinking things around health insurance, labor law, et cetera, to shift from a group setting as the norm to individual self employed entrepreneurs?
- [15:44]
A
Speaker Gingrich, you hit the nail on the head. And I've been writing a lot on exactly this point. Because even if this shift is gradual to more people becoming independent entrepreneurs, self employed in independent work context, it matters a lot because our safety net and benefit systems all assume a stable W2 employment. And so if more people earn income across multiple clients or projects, employer tied benefits and systems make less sense in that world. And I wanted to highlight that we're seeing a momentum today for what's called portable benefit systems. And this is happening both at the state and federal level. It's an institutional update to match a more flexible and fluid workforce. And the key idea there is think like HSAs, but for everything, or those sort of systems where benefits are worker owned, they travel with the worker, they're not tied to a particular job or company, and they move with the worker as they go from client to client or job to job. And I think that is something that we really need to be thinking about for the 21st century, for how to modernize this institutional system that's been around for a while and that is based on a 1930s design of the workplace. Right. So in that setting you have one person who's a W2 employee, stays at one company or one career for their whole lives into a much more fluid and flexible workforce that we see today. And I wanted to highlight, even with young people today. So post Covid, we're looking at surveys of young people entering into the job market. A lot more people today than 10 years ago are going into independent work. And I don't mean kind of Uber and Lyft sort of thing. I mean, like they say, I'm going to be an entrepreneur and start my own YouTube channel and that's my first job out of college. And so that is all changing now. And we really need to rethink kind of the institutional framework that we have where everything is tied to a W2 job. And I think that's why portal benefits reforms and systems bills introduced at the state and federal level today are helping to create an update to match this modern workforce.
- [18:08]
B
There are streaming channels now in which people make amazing amounts of money.
- [18:14]
A
A lot more than an entry level job.
- [18:18]
B
Less studying and more talking and posturing. Maybe the key to a higher income level.
- [18:25]
A
But I think you're right and I just want to say the data and the research 100% supports that. We are seeing a lot more people at younger ages go into these independent work jobs, influencers, YouTube channels and so forth. Now, I don't know if that's a good thing or a bad thing and what it means, but I do think it is forecasting what's going to happen in the future. Because AI is going to take that and multiply it by 10 according to my analysis on the ways that AI will change how work is organized. And that's why it's really important to get the policies and the institutional framework right for a workforce 10, 15 years from now, where a lot more people are in these independent work or influencer or other self employed work arrangements. Because again, our institutional framework is designed around W2 jobs, an employee employer relationship. Great. A lot of things are designed around that. So we really need to be talking about the infrastructure today.
- [19:24]
B
It makes me think that Congress ought to be holding hearings on the rise of entrepreneurial systems that are individual structures. You know, what's the health insurance, what's the taxes, what kind of records do you need, et cetera, which will be totally different. And most of that will be mediated by artificial intelligence. So you actually will have an ability to be your own bookkeeper on a scale that would historically have been impossible.
- [19:58]
A
Yeah, that's right. And I couldn't agree more with you, Speaker Gingrich on Congress needs to be talking about these. And actually they've started a little bit with these hearings. I've actually testified at four different hearings trying to unpack independent workforce and portable benefits. Sometimes the problem that you get is this turns into an independent contractor issue with people bringing up ideas of like, well, this is about employer misclassification of workers. And then we're sort of missing the boat at that point. We're like, no, we're talking about independent entrepreneurs and real freelancers and self employed workers. And so we need to be thinking about those people, not just about, oh well, this is an independent contractor in a vulnerable position and the company is exploiting them or misclassifying them. So unfortunately those conversations, some of them have started, but they keep being dragged into some of these older debates about independent contracting. Rather than thinking forward about the new type of jobs that are emerging today and tomorrow.
- [21:14]
B
Won't there be a great deal of tension and confusion about the rise of these brand new sort of self supported, independently functioning entrepreneurs?
- [21:26]
A
Yeah, that's a great question. And we're actually seeing it today. There are bills being introduced at the federal level and at the state level that enable portable benefit systems or try to modernize labor laws and health insurance systems to create benefits that are more portable. We are seeing unions today come speak against those. And their argument is that look, most of these workers are misclassified or they're exploited. And again, they keep going back to this issue of independent contractors and so forth. I understand where they're coming from, but the real question is that 10 years from now a lot more workers are going to be entrepreneurs and self employed. And how are unions going to operate in that framework in that modern economy? And I think it's really important to try to think about how can the infrastructure around labor unions also be reformed so that they're not kind of missing the boat and still living in the 20th century with a 21st century economy. I've written a little bit about how we can separate the good aspects of unions, which is worker voice, which is essential, and separate that from our labor structures around unions that create monopoly kind of bargaining structures and so forth. So, kind of monopoly bad, worker voice good. How can we create a system that allows for worker voice aspect to thrive, but how do we move away from the monopoly bargaining structures that kind of bring things down?
- [23:01]
B
When you look at all this, it's a huge level of change. If AI is the future, what's the role of education and all that? And should we be rethinking how children get educated if they're going to be living in a world of artificial intelligence?
- [23:17]
A
You could go two ways with this. So it's interesting that China is taking an approach towards AI where they're basically formally requiring educational systems to start training kids as young as six years old to start using AI, because I think they're predicting a future where this is the most valuable skill and tool that you can have. So there's an argument for the China model, some of the research and literature I've been reading for the other side of the model is that no, we should not introduce AI to children this young because they need to be able to develop these critical thinking skills and other foundational skills early on so that they can later be able to assess AI's work, if that makes sense. And so there are these two different models I've seen being discussed about educational systems with AI. And again, the second model is more like don't introduce AI early, let them build the skills independently as humans to critically assess and so forth. And then the AI can augment those skills. So I can't tell you which one is the right one. Off the top of my head. I have two young children, 2 and 4 years old, and I'm thinking about this as well. And I was like, I would like them not to use AI early on and then train on it much later in the way that a lot of us. I grew up going to school in 1990s and so forth. And I think we started to take on computer skills slowly, but it was still really helpful. It was great. I remember it being introduced in middle school, but I think AI is a whole nother ball game because it can make you maybe think in a more lazy way because you're saying, oh, I'm just going to give this to AI versus if you can't use AI, it might force you to kind of unlock those critical thinking skills and dig deeper. So it'll be really interesting to see how this unfolds in the educational debate.
- [25:17]
B
I notice going to restaurants you'll have four, five, six year old kids have a phone and they're playing games. It's a different world. You are going to be very busy because you're a serious student of an area that is going to be changing so dramatically and in so many different unpredictable ways that just keeping up with it is going to keep you pretty busy, don't you think?
- [25:45]
A
Yeah, I think labor markets are the most important things right now in terms of economics and policy. We're also at a really important moment in American labor. As you can see, there's this new sort of coalition on pro worker that includes a lot of Republicans and Democrats and they're pushing on this. We're pro worker now rather than pro business. And so it's a really fascinating time to be working on this because it's harder to say, maybe 20 years ago or 10 years ago you might say like, oh, the Republicans won't support unions. But now that dynamic is changing a bit because there's this coalition of we're all pro worker, we don't want to be pro business anymore, and so forth. So it's a really interesting time. But what doesn't change, I should say, is trade offs, right? So the same trade offs will be there. If you have significant wage gain increases, new rigid rules in a way that is not offset by increases in productivity, you're going to get the same dynamics in the future, which is slower employment growth, layoffs, more pushes towards automation. We've done research to review 147 economic studies looking at union monopoly structures and bargaining power. And what I can report consistently is that you get these short term gains mostly on wages and benefits, but you also get long term costs, slower employment growth, more downsizing, more closures and so forth. And people don't think about those long term costs. They just want to think about the short term wage gains.
- [27:21]
B
I want to thank you for joining me. You're in a field that is going to change constantly and it's going to be very important. And our listeners can learn more about the work you're doing if they'll visit Mercatus.org so thank you very, very much.
- [27:35]
A
Thank you so much for your time.
- [27:39]
B
Thank you to my guest, Leah Palagashvili. New 12 is produced by Gingrich360 and iHeartMedia. Our executive producer is Garnesy Sloan. Our researcher is Rachel Peterson. The artwork for the show was created by Steve Penley. Special thanks to the team at Gingrich360. If you've been enjoying Newts World, I hope you'll go to Apple Podcast and both rate us with five stars and give us a review so others can learn what it's all about. Join me on substack@gingrich360.net I'm Newt Gingrich. This is Newts World.
- [28:27]
A
This is an iHeart podcast. Guaranteed Human.