STEVE INSKEEP, HOST:
Let's follow up on a warning about our future. Pete Buttigieg, the former and possibly future presidential contender, was on NPR last week. And during our video interview, he warned that Americans are not preparing well enough for the economic dislocations that may come from artificial intelligence. Let's listen.
(SOUNDBITE OF ARCHIVED NPR CONTENT)
PETE BUTTIGIEG: The economic implications are the ones that I think could be the most disruptive the most quickly. We're talking about whole categories of jobs where not in 30 or 40 years, but in three or four, half of the entry level jobs might not be there. It'll be a bit like what I lived through as a kid in the industrial Midwest when trade and automation sucked away a lot of the auto jobs in the '90s, but 10 times, maybe 100 times more disruptive.
INSKEEP: So what do we do about that? We have one of many views on this issue. It comes from Erik Brynjolfsson, who is at the Stanford Institute for Human-Centered AI and has studied this issue for many years. Welcome to the program.
ERIK BRYNJOLFSSON: Good to be here.
INSKEEP: First, is Pete Buttigieg right to raise this concern about the very near future?
BRYNJOLFSSON: Yeah, he's spot on. We are seeing enormous advances in the core technology and very little attention being paid to how we can adapt our economy and be ready for those changes.
INSKEEP: What are the jobs that are going away partly or entirely, do you think?
BRYNJOLFSSON: Well, just to be clear, there's transition in both directions. There are jobs that are disappearing and there are new jobs being created. So some of the jobs that are changing the most are in coding and software engineering and in call centers, which I've studied. And in both of those areas, we're already beginning to see some effects, especially on entry-level jobs.
INSKEEP: I'm also thinking of an auto factory that I visited earlier this year in China. And they had some human employees, but robots were doing an awful lot of the work. And I would imagine robots might do more and more of it.
BRYNJOLFSSON: Absolutely. No, I visited one of those factories over in Shenzhen and I saw the same thing. They're all going to trend in the direction of more automation. That's just more cost effective. And the people at the factories were telling me it leads to better quality and more consistency as well.
INSKEEP: Well, that can be. But of course, there's also the dislocation. So let's focus on that for a second. Pete Buttigieg made that comparison to the industrial Midwest, where there were entire communities that were devastated. Could we see that kind of effect?
BRYNJOLFSSON: I think it's a very apt analogy. And the ideal thing is that you find ways of compensating people and managing a transition. Sad to say, with trade, we didn't do a very good job of that. A lot of people got left behind. It would be a catastrophe if we made a similar mistake with technology that also is going to create enormous amounts of wealth. But it's not going to affect everyone evenly, and we have to make sure that people manage that transition.
INSKEEP: Give me some ideas. How would we shape our future in a more positive way?
BRYNJOLFSSON: Well, the first thing a lot of economists, including me, would go to is worker retraining and understanding what are the kinds of skills and tasks that are going to become more important going forward. And so a lot of interpersonal skills, a lot of management skills. I think as we create these agents, we're all going to become CEOs of our own little fleet of agents. And learning those management skills that apply not just to humans but increasingly to agents is something that can be taught. And if we each have our own fleet of agents, then we can each be more productive.
INSKEEP: You better explain what an agent is for people who aren't familiar.
BRYNJOLFSSON: Sure. AI, we're all using LLMs increasingly and using them to give us bits of text back to us and give us some advice. Agent is taking that one step further, where the large language model doesn't simply give us some text back but actually takes an action, goes ahead and makes an airline reservation or buys something for us or carries out some other task.
INSKEEP: I want to add a note of skepticism about retraining. I understand broadly why it would be a good idea. But I think about the industrial Midwest and the way that people talked about job retraining as old industrial jobs went away. And some of the effects of that included job retraining programs, which sound good, but didn't really help people very much. An emphasis on education, which sounds great, but maybe it just increased the price and the demand for education. And not everybody ended up with a very fruitful job on the other end. This is hard.
BRYNJOLFSSON: It's hard. But, you know, there's still a wage premium for people who are more educated versus less educated. And it doesn't necessarily mean retraining to do more of the same. It's often moving into new industries, new kinds of tasks. I do think that interacting with humans face-to-face is something that a lot of people prefer. And that's something that, you know, by definition can only be done by other humans.
INSKEEP: Well, as you say that, I'm wondering if we have the wrong emphasis on education. We've been trying so hard to bring up STEM fields, science and math and so forth, engineering. Some of those jobs can be replaced. And maybe we should've been emphasizing more of the liberal arts, which deal with humanity.
BRYNJOLFSSON: I think it's a both-and. I absolutely do think there could be a revenge of a lot of humanity tasks. And there may be opportunities there for more creative, artistic, interpersonal work. At the same time, there are huge premiums for certain kinds of engineering, math, science, coding. I mean, you may have seen some of the outsized salaries that are being offered to people who are really good at creating AI.
INSKEEP: Yeah, absolutely true. Do you worry that the creative jobs could go away, too, however? I think about the movie industry, where there's already been a lot of tension and conflict about the idea that maybe you don't even need the actor to be there. The AI can make it up.
BRYNJOLFSSON: It's amazing what's happening in that industry. And I think, again, that's a place where the jobs are going to change a lot. I do think there'll be opportunities for people like you and me to make our own fun little movies. I've been using ChatGPT to write, you know, poems to my wife and songs that I couldn't have done previously. Maybe next year I'll be making movies.
INSKEEP: Let me stop you there because somebody's going to be listening and saying, wait a minute, you asked ChatGPT to write a poem to your wife?
BRYNJOLFSSON: Full disclosure, I did tell her where it came from. But somehow, she still appreciated it.
INSKEEP: She didn't (laughter) want something a little more from your heart?
BRYNJOLFSSON: It was so funny. You would've appreciated it, Steve. Nobody would've not thought it was funny, what I wrote for her.
INSKEEP: (Laughter).
BRYNJOLFSSON: With help.
INSKEEP: You're saying it's definitely funny, is what you're saying (laughter)?
BRYNJOLFSSON: It was definitely funny because some things are just funny, even if you admit that AI helped you a little bit.
INSKEEP: Granting that you seem basically optimistic, is there something that keeps you up at night?
BRYNJOLFSSON: Oh, definitely. I'm optimistic about the potential to create a lot more wealth and productivity. I think we're going to have much higher productivity growth. At the same time, there's no guarantee all that wealth and productivity is going to be evenly shared. We are investing so much in driving the capabilities for hundreds of billions of dollars, and we're investing very little in thinking about how we make sure that leads to widely shared prosperity. That should be the agenda for the next few years.
INSKEEP: Erik Brynjolfsson is a senior fellow at the Stanford Institute for Human-Centered AI and also directs the Stanford Digital Economy Lab. Thanks for your insights.
BRYNJOLFSSON: My pleasure.
(SOUNDBITE OF B. FLEISCHMANN'S "PASS BY") 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.
