SCOTT SIMON, HOST:
Two days after the midterm elections, the tech-heavy NASDAQ lost over $2 trillion in market value. Tech stocks rebounded modestly on Friday. That market chaos was over worries of an AI slowdown once Democrats take control of Congress. Daron Acemoglu teaches at the Massachusetts Institute of Technology, and he received the Nobel Prize in economics two years ago. We asked the professor how Congress might lower costs for Americans, and he said, that's the wrong question, considering AI's impact on the workforce.
DARON ACEMOGLU: So the question is, how can we create more good jobs for people of different skills? - not just those with degrees in computer science or entrepreneurs and managers, but the typical working family getting their salary in blue-collar or white-collar jobs. How can we do that? That's the question.
SIMON: So you believe the approach needs to be not lowering prices, but increasing wages?
ACEMOGLU: Increasing wages, yes, absolutely. I think, you know, of course, nobody likes going to the grocery store and seeing that whatever they used to pay for milk, now they have to pay 20% more. But that's a much less of a burden when you see your own salary increase at the same rate or at a higher rate. But it's not an easy feat.
SIMON: Like, what could we do about it?
ACEMOGLU: We impose a lot of taxes on employers and employees. In other words, we tax labor income quite heavily. On the other hand, we don't tax capital income, which typically goes to the wealthiest people in the country. So creating a fairer, more symmetric tax code will help. But most important challenge facing jobs is artificial intelligence. Many economists and technologists expect the rate of change in artificial intelligence improvements to continue. And as this happens, more and more jobs will be taken by artificial intelligence.
So that means we have a conundrum. When we need to create more jobs for people, create more wage growth opportunities for people, technology is going to make that harder. So that's why I would say we have to create an agenda of steering AI in a more pro-worker direction so that rather than automate work, we get AI to actually help employers to find ways of using workers more productively, and that will be a pathway for encouraging firms to create more jobs.
SIMON: Isn't the whole idea, or at least a significant part of the idea, of AI to reduce the number of jobs, though?
ACEMOGLU: Of course, employers are going to go down the path of saving labor costs, meaning automating work. But when you look at AI, it is a very versatile technology. It can provide information. It can expand what humans do. It enables workers to be much better at solving problems. It enables them to gain new expertise. And as it does this, it could also create new jobs and help wage growth. But this won't happen by itself. It won't happen when the majority of the investments go into the dream of creating more AGI, more superintelligence and sidelining workers.
SIMON: How quickly could this be done?
ACEMOGLU: It can't be done that quickly. That's the problem. But I think people are also reasonable. If they see that efforts are being made to change the long-term trajectory of the economy and that's communicated to them, I think they will be much more open-minded about this. I think the problem with AI right now is that it is secret. It is happening behind closed doors. And, of course, there is no democratic voice providing an input into where AI should go. Trump administration has taken a completely laissez-faire approach, and if that continues, I think the current direction is going to get cemented. Then it's going to be much harder to reverse it. I'm not saying that it's easy to reverse at the moment, but there are ample opportunities to use AI in very different ways than this AGI dream.
SIMON: It's hard not to notice - you, of course, are a Nobel laureate in economics, and we want you to talk about economics. But you are emphasizing for us that the future of almost everything we're talking about is tied up in artificial intelligence.
ACEMOGLU: Unfortunately, it is. When you have a promising technology, plus, you put so much investment in it, of course it's going to shape every aspect of our society. But shape is the right word here, not improve. There is this rhetoric that comes from some AI enthusiasts and perhaps some policymakers as well that just let AI rip, and everything's going to be great. We're all going to benefit from it. Technology always creates winners and losers, and policy institutions and aspirations shape who these losers are and who the winners are, and how can we make more winners and less losers, when something is that important, to say, that a handful of companies who decide its entire future with no democratic input cannot be the right approach. So this is part of economics. Economics has to deal with these transformative technologies as one of the most exciting parts of economics.
SIMON: Is our democracy up to it?
ACEMOGLU: Well, I think it's definitely not up to it if we let AI rip and more and more inequality build up and a couple of companies become so large. That kind of inequality, that kind of control over information by a handful of companies, I don't think our democracy would survive. But right now, our democracy is strong. I think the midterms show that. And we should use that strength in order to use that democratic muscle to also shape AI in a way that's socially beneficial, in a way that democratic voice supports.
SIMON: Daron Acemoglu, professor of economics at MIT. Thanks so much for being with us, sir.
ACEMOGLU: Thanks for talking to me.
(SOUNDBITE OF HLFMN'S "GANGS OF CATS") Transcript provided by NPR, Copyright NPR.
NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.
