Transcript
ANNOUNCER: NPR.
[COIN SPINNING]
ADRIAN MA: This is The Indicator from Planet Money. I’m Adrian Ma. Last week, we held a live virtual event called Ask an Economist. This is a new Q&A format we’ve been trying where you listeners get to pose your burning questions to economists about hot topics in the news. The theme for this event was Will AI Take My Job? And we were joined by a few hundred NPR supporters who have lots of questions and feelings to share on the topic.
SPEAKER 1: How do you incentivize the people developing AI, like Anthropic or OpenAI, so that they know how to be pro-human?
SPEAKER 2: If people are satisfied with AI-written books, are they ever going to buy mine?
SPEAKER 3: Where’s that line between providing my family with a good life now versus just saving everything for a possible layoff?
MA: Will AI bring on a white collar job-pocalypse, or could it usher in a new golden era of work? For answers, we were also joined by two influential economists who are shaping how we’re thinking about this right now– Daron Acemoglu, a Nobel Prize-winning professor from MIT–
DARON ACEMOGLU: A lot of jobs may disappear, and it is also up for grabs whether there will be new jobs of the same quality.
MA: –and Erik Brynjolfsson, director of the Stanford Digital Economy Lab.
ERIK BRYNJOLFSSON: We can steer AI in ways that are more likely to complement humans and therefore create new opportunities.
MA: Today on the show, we’re sharing a few highlights from this fascinating hour-long conversation. And after the break, our experts will get into the potential and the pitfalls of AI in the workforce.
MA: So let’s start with some good news. Stanford’s Erik Brynjolfsson told us that there is currently no evidence that AI is causing widespread job displacement. But as we heard throughout the event, people’s fear that AI could take their jobs is real. Are you there, Ben?
BEN: Yes, hi.
MA: Ben, go ahead. What’s your question? Ben joined us from Seattle. He works as a principal software engineer at a big tech firm. He’s a father of three. And Ben described this dilemma that he faces. He said it feels like AI is hanging over his head like a sword.
BEN: Where’s that line between providing my family with a good life now versus just going completely paranoid and cutting all optional spending and just saving everything for a possible layoff, especially, you know, when I think I would be lucky to ever be hired again as a software engineer, and certainly not at the level that I’ve reached after almost 20 years? You know, it’s very stressful. And I worry about people coming into this industry, but, you know, I worry for my family as well. Where’s– how do you think about that?
MA: Ben, thanks for your question.
BRYNJOLFSSON: Well, I do think we’re entering probably a very turbulent period. And it’s very hard to predict exactly what’s going to happen. The technology is getting immensely powerful, and it’s going to have a lot of unpredictable effects. And so I share your concern, and I sympathize with that. It’s tough.
MA: Now, when it comes to entry-level jobs, Erik told us he and his colleagues have seen warning signs in their research. In one study, they found that since ChatGPT and other AI models began rolling out, 22 to 25-year-olds in jobs like customer service and software engineering, they’ve had a harder time getting work. Their employment levels were almost 20% below their peers in less exposed jobs. However, Erik says, more senior workers like Ben, our software engineer, seem to be doing better.
BRYNJOLFSSON: People who could define what the project was and often would delegate parts of it– there’s the front engineer, the back end, the database, et cetera– the different parts of it had to be done by more junior people. Now, agents are doing a lot of that, as I’m sure Ben is– probably has a little fleet of agents that he’s managing. And so that could be something where there’s growing demand for that end of the market. Now, I can’t say, a year, three years, five years from now, how that’s going to evolve. The technology is moving very rapidly. So I do think it’s wise to have some caution and have some reserves in case things change in an adverse way.
MA: As AI gets better, certain jobs could be automated out of existence. But as our other guest, MIT’s Daron Acemoglu, argued during the event, there is another way. Daron talked about what he calls “pro-worker AI.” And he gave us an example of how AI could support educators in their work.
ACEMOGLU: A lot of new technologies in the ed space are aimed at replacing teachers, whereas I think there are a lot of pro-teacher AI opportunities where you give better tools to teachers, but teachers still remain central for the learning experience of students. But they can individualize the teaching experience much more effectively and at low cost in a way that wasn’t possible before. And that expands teacher expertise, expands teacher capabilities, enables them to do– perform new tasks, doesn’t replace them, doesn’t sideline them. So there are many, many other examples, from nurses to journalists, to creative access, white office– white collar or– work, blue collar work. I think there are a lot of things of this sort.
MA: So this is painting, like, what potential positive things could be done with AI to augment workers. But I think you would argue that this is definitely not the track that AI companies are on.
ACEMOGLU: This is not a track. Again, just like we look at what they’re saying, they’re saying AGI is the burning desire. AGI is where they’re aiming. AGI is just around the corner. That’s just a code word for saying, make the models more and more human-like.
MA: AGI, meaning Artificial General Intelligence that has human-level capability. Now, Erik says he’s not quite as convinced as Daron that the frontier AI labs are focused on replacing humans, but he said that the technology still needs to be steered to more broadly complement humans. One way to do that, he said, was to have better benchmarks and metrics.
BRYNJOLFSSON: It is remarkable how good a lot of these engineers are at hitting a benchmark once you put it out there. And we have something called the Stanford AI Index, which is 420 pages filled with all of these benchmarks. Almost all of them are how well can a machine, by itself, with no human help, solve a problem. It’s considered cheating if the human helps. I think that’s the wrong mindset. We should be building benchmarks where humans and machines are encouraged to work together, and the tool designers are encouraged to make tools that work well with humans, rather without humans. Right now, one of the frustrations, for instance, of radiologists is, they have these black box imaging machines that recognize medical images, but they can’t explain what they’re doing very well or how they come to their conclusions. It would be better if they were designed from the beginning to work with the humans.
MA: Erik and Daron also weighed in on how AI might affect other jobs, including paralegals, game designers, and novelists. But for now, we’ll end on their advice for those of you just starting out. OK, so in the– I guess, the minute that we have left or so, Kara asks, how do we speak to college-aged students and families about their fears regarding the worth of their degree in the future? Eleanor had a similar question about what advice you would have for college students planning their careers.
ACEMOGLU: I mean, I would say, the flexibility is very important. The general principles that Erik and I talked about, it’s much better to work alongside AI than compete against AI. But also, I think the young generation, the new generation, has to learn to shoulder the responsibility of being– becoming democratic citizens in a world of AI. And I think that’s a very important thing. We’re not passive observers here. It is our votes, our decisions, our voice that can influence the future of AI. It’s not just the whims of people in frontier labs.
BRYNJOLFSSON: Big plus one to that. When people hear AI, they should be thinking, Amplified Intention, that we actually have more power to change what we’re doing. And these tools are– since they are so incredibly powerful, people who have good ideas, they’re going to be able to implement them much better than they did before. And I think it could be a very exciting time for a young person to be entering the workforce who’s willing to dive in and have that kind of agency.
[MUSIC PLAYING]
MA: Thanks, again, to Daron Acemoglu and Erik Brynjolfsson and all the NPR supporters who turned out with their questions and stories. To hear this entire event or come to the next Ask an Economist, make sure you’re signed up for NPR+ via plus.npr.org. This episode was produced by Viet Le and Angel Carreras. It was edited by Julia Ritchey and fact-checked by Cooper Katz McKim. It was engineered by Kwesi Lee. Kate Concannon is our show’s editor. And The Indicator is a production of NPR.
Copyright © 2026 NPR. All rights reserved. Visit our website terms of use and permissions pages at www.npr.org for further information.
Accuracy and availability of transcripts for NPR programming may vary. Transcript text may also be revised to correct errors or match updates to audio. Audio may be edited after its original broadcast or publication. The authoritative record of NPR’s programming is the audio record.
Please note that NPR platforms and products also include some programming produced by third parties; NPR cannot guarantee transcripts for that content and some transcripts may be produced using automated or other tools.