1. Podcasts
  2. 1A
  3. Pending: AI In The Workplace
1A

10 Nov 2026 · 44m · 16 chapters

Episode artwork

Pending: AI In The Workplace

In short

The episode “Pending: AI In The Workplace” explores uncertainty about how AI will affect jobs, hiring, and day-to-day work. It opens with listener concerns (e.g., hiding AI use in academia) and cites a Pew finding: 52% of respondents are more concerned than excited, and under-30s are more concerned than excited for the first time in five years. Guest Erica McIntyre (Stanford IER; previously BLS Commissioner) argues AI’s labor-market impact is still early: only about 1 in 5 firms have adopted AI, often in “shallow” functions like customer service and internal help desks. She notes hiring declines for 22–25-year-olds in “AI-exposed” occupations after ChatGPT’s 2022 launch, but says it’s hard to isolate AI from pandemic and remote-work effects. Guest Arun Sundarajan (NYU Stern; studies AI/future work) and LinkedIn chief economist Liz Wilkie (LinkedIn) discuss generative AI being strong at ideation and option generation but weaker at verification and secure complex coding. They cite LinkedIn data: AI literacy skills on profiles are growing, and job postings requesting LLM/AI tool skills rose 65% year over year. Notable examples from listeners include AI summarizing legal depositions (40 hours to 45 minutes), AI-driven layoffs after training tools, and burnout from poorly designed AI programs.

Key claims

AI may reduce entry-level pipelines and freelance work (not fully captured in job stats), while workers can improve prospects by verifying outputs and combining AI with domain judgment and soft skills.

Written by AI. May contain mistakes.

Chapters

0:23

AI's Growing Role in the Workplace

Explore the increasing integration of AI in work environments and public sentiment.

“It's in your email when you start composing a message to a colleague.”
3:10

Seasonal Changes and LifeKit

Hear about LifeKit podcast episodes offering lifestyle tips for the fall season.

“Yes, the days are getting shorter and colder, and the rain can be a drag.”
3:33

AI Impact on Labor Market with Erica McIntarfer

Discussion with Erica McIntarfer about AI's effects on job markets and employment data.

“And let's start with a picture of what's happening in the labor market now.”
14:30

AI's Impact on Work Dynamics

Discussion on how AI is being integrated into workplace communication.

“We're talking about artificial intelligence and how it's changing how we work and the careers we can have.”
15:12

Guest Introductions

Introduction of Arun Sundarajan and Liz Wilkie, experts on AI and the workplace.

“And a member of the 1A Text Club writes, as a technical writer for software development, we do use AI.”
15:44

Generative AI Tools: Strengths and Limitations

Exploration of tasks AI tools excel in and where they fall short.

“What I'm seeing is that they are really good for tasks where you need to generate lots of options.”
17:06

Evolution of AI Skills in Job Market

Examining changes in job listings related to AI skills over time.

“But as Erica was pointing out, you know, there's the capabilities that AI has are very different from it's deployed, what it's being deployed at.”
18:36

AI's Structural Impact on Hiring

Discussion on the current hiring landscape and AI's influence on job availability.

“So we're seeing a response on both the supply and the demand side for these skills.”
19:46

Future of Entry-Level Jobs with AI

Concerns about the impact of AI on entry-level opportunities and human capital.

“You know, certainly AI is not showing up as mass layoffs, but it may be showing up as, you know, jobs that never get offered.”
22:40

Challenges for Senior Workers with AI

How AI skills affect the responsibilities of senior workers and workplace dynamics.

“And this will then lead us to lean excessively on AI, not because it's better, but because we don't have an alternative.”
Show all 16 chapters
24:42

Companies Experimenting with AI

Observation of companies navigating uncertainty while integrating AI.

“What we also find is that young workers in particular with AI skills do have better hiring outcomes.”
26:56

Real-World AI Application Challenges

Discussion on the practical challenges companies face when implementing AI.

“It takes a few years of experimentation, of figuring out how to do work differently before the gains start to come.”
28:06

The Tentative Job Market and AI's Impact

Explore how AI is influencing entry-level jobs and company strategies.

“And so part of the tentativeness that we may be seeing in the job market at the door, not the exit, right?”
31:55

Worker Adaptation to AI Tools

Learn how workers are leveraging AI tools to enhance productivity and the implications for job security.

“Melissa emails, I work as a law clerk summarizing thousand page depositions for my law firm.”
42:00

The Inevitable Adoption of AI

Explore the inevitability of AI adoption in various industries and its implications for workers.

“And the people who know how to adapt to the new needs that are going to be demanded from humans as AI starts to take over some of the things that humans used to do.”
45:49

Challenges and Burnout in the AI Era

Discuss the potential for burnout among workers as AI is integrated into workflows and the ongoing experimentation in this space.

“I think that that's a really understandable reaction.”
A man walking through an airport, asking his phone a question

Ask about this episode

Get the gist, explore an idea or ask about something said.

Claude

On 1A, what are the key points of this episode?

Searching your listening

Transcript

Automatic transcript. May contain errors.

0:00This message comes from Rinse. With Rinse, your laundry will come back fresh, folded, and ready to wear right on schedule. One tap or text is all it takes. No calling, no reminders. Same as always. Rinse. It's time to be great. More at rinse.com.

0:23It's in your email when you start composing a message to a colleague. It's in your search results offering an answer to your question. It can help you write a report, design an image, even create a full video. The capabilities of artificial intelligence tools have increased significantly in the past four years. Now they're becoming a part of the workplace, either because workers want to use them or because company executives see value in adding them to their business. We surveyed you in the 1A Tech Club asking if you use AI in your work. One of you wrote, yes, but I feel like I have to hide it because I'm in academia.

0:56And Angel responded, I see AI as a sophisticated research librarian with enormous computational capabilities, but one that you have to question and independently verify. My approach begins with my thinking, not AI's. As generative AI becomes much more common in our work, people's opinions of the technology is also changing. A Pew survey from this summer found a majority of respondents, 52%, are more concerned than they are excited about the technology's role in daily life. And for the first time in five years of asking the question, a majority of people surveyed under 30 years old said they were more concerned than excited.

1:36How much of this worry and attention on AI and work is based in reality? How are employees changing their work to include AI? What do we know and not know about what this technology is doing and will do to our careers? I'm Jen White. You're listening to the 1A Podcast. Today, uncertainty in the job market and what artificial intelligence means for workers now and in the future. It's part of Pending, our series looking at uncertainty and how it shows up in modern life. We'll be back with more after this short break. Stay with us.

2:11Stay tuned for the following announcements and messages from our sponsors. This message comes from the UPS Store. The UPS Store is open this holiday to help make sure your small business can stay at its best. During the holiday, customers have access to professional printing, packing and shipping, shredding, mailbox services, notarizing services, and everything else your small business could need. Find nearby locations and keep tackling important tasks at the upsstore.com. Most locations are independently owned. Products, services, pricing, and hours of operation may vary. See Center for details.

2:45Making Up First takes overnight shifts from dedicated journalists across the country, and it's only possible with listener support. When you give a few bucks a month through NPR Plus, you power our newsroom. And as a thank you for your support, you can save over seven hours a year with sponsor-free episodes. Give a little and get more NPR all in less time. Sign up right now at plus.npr.org. Happy fall! Yes, the days are getting shorter and colder, and the rain can be a drag. But this is also the perfect time of year to get cozy. LifeKit has lots of ideas so you can embrace the change of seasons, with episodes on everything from making soup to exercising outdoors.

3:27Listen to the LifeKit podcast in the NPR app or wherever you get your podcasts. Welcome back to the 1A podcast. We're discussing AI and work. And let's start with a picture of what's happening in the labor market now. Joining us from Palo Alto, California, is Erica McIntarfer. She's a senior policy fellow at the Stanford Institute for Economic Policy Research. From 2024 until August 2025, she was the Commissioner of the Bureau of Labor Statistics. Erica, welcome to 1A. Hello. Now, Erica, you study employment data and other indicators of the health of the labor market. When someone asks you how AI impacts jobs, what data do you look at?

4:08So this is one of the most common questions I got both when I ran BLS and when I came to Stanford because there's just been a lot of anxiety about AI's impact, particularly for white-collar workers. So I look at a lot of data when trying to assess AI's impact, employment data, unemployment data, firm data on how firms are adopting AI. And generally, the picture that comes across when you look across this swath of data is of a technological revolution that is still very much in early stages. So not much impact, very little impact on employment, on unemployment. Firm adoption is not maybe as far along as many people think.

5:04Only one in five firms have adopted AI. In a brief you published in July with colleagues at Stanford, you noted a higher level of unemployment among new graduates and workers in the early stages of their career. And you and your colleagues say artificial intelligence could be a factor there. What are you seeing specifically in the data for those workers? Yes. So there has been a fall off in hiring of young workers in what's commonly referred to as AI exposed occupations. And these are jobs where more of the tasks can be automated with AI. So examples here are like software developers and customer service agents, where AI capabilities are pretty good.

5:54And when you look at some microdata, you do see evidence that hiring of 22 to 25-year-olds has fallen off since approximately around the launch of ChatGP in 2022. And so this is one of the few signs that AI is disrupting the labor market. The complication is it's really hard to know how much of this is due to AI. So, unfortunately, the fall of 2022 is a very tough time to isolate effects on the labor market. There was a lot going on. We had just come out of the pandemic. There had been a lot of pandemic overhiring in a lot of the industries where you see these declines in hiring. The Fed had begun hiking rates.

6:46There had been an enormous pivot to remote work, which makes it hard to train young people. So maybe firms were hiring fewer young people because it was harder to train them when much of their staff was remote. But it's important to note this is really a substitution away from young people towards hiring more experienced people, which is one of the reasons we're not seeing it in the aggregate data. You mentioned software developers, and that's one profession that shows up in the news quite often in the context of AI disrupting work. We've also heard stories recently about mathematicians seeing a disruption.

7:22the customer service industry seeing a disruption. What professions have you looked at or are you watching most closely to see if a change in hiring could be related to AI? So I think the software developers are a great example of an occupation to watch very closely because it's going to give us maybe an early look at what disruption from AI could look like for a couple of reasons. One is software developers are highly concentrated in industries where adoption has been high. So as I mentioned, firm adoption, only one in five firms are adopting AI, but they're highly concentrated in information-intensive sectors, so where software developers work.

8:10So tech, finance, banking, law firms, professional services. The other reason is because AI capabilities are what's called jagged. AI is not, as many people who've used AI have discovered, AI is not perfectly good at everything. But it is very good at coding. And so this is a domain in which AI capabilities are very high. So we have software developers sit in this unique space where they're both in an industry that is being, that is adopting the technology at a high rate and also the technology is transforming their jobs. So right now, software developer jobs are changing a lot. Like what they're doing from day to day is changing, but we're not seeing it show up in employment statistics.

8:57Employment for this occupation remains, the growth in this occupation remains fairly steady. So when you say that only one in five companies are using AI in their work, What does that mean from a practical perspective? What are these businesses saying about how they're using it? So that's a really good question because one of the reasons we are likely not seeing enormous disruption in the overall labor market from AI is while 20 % of firms, one in five, are using AI, their use of AI is what you would describe as shallow. So it's only a handful of business functions typically. So the most common ones are customer service, internal help desks for HR and IT.

9:49So it's not a deep transformative use of technology yet. And also where most firms are is they're still experimenting with how to use AI. So a lot of these experiments are still in pilot stages and have not actually been deployed throughout the enterprise. So I'm really curious to hear from you how you think about this. I think it's fair to call it a disconnect between what the data says and the way people feel about artificial intelligence in the workplace right now and what it could mean for their future job prospects. So there is this disconnect. And one reason for it is that where the model capabilities are in terms of what they can do is so far quite advanced of what businesses are actually using AI for.

10:48So there's this big gap between the potential use of the technology, what it's capable of, and how businesses have been able to figure out how to use it profitably. And that gap might take a while to close. Model capabilities in AI improved enormously fast. Businesses are typically slower moving organizations and their adoption is moving along at more, you might call a human pace. So if someone listening right now is worried that no matter what's happening in this moment, artificial intelligence tools are going to rapidly change, you know, hiring and firing the tasks that office workers do. What would you tell them about the overall picture this data presents right now?

11:40So the picture that the data presents is one where AI is transforming the kinds of tasks that you do at your job and maybe how you do them. So when you survey workers, they do report using AI at work. I will say most of them do not report using AI very extensively. Software developers here are an exception. Software developers do use a lot of AI right now. But for your average white-collar worker, they might use AI like less than an hour a day to help them write emails, to help them research something. Their use is, again, pretty shallow. It's not yet transforming their job. That's Erica McIntyre for her.

12:25She's a senior policy fellow at the Stanford Institute for Economic Policy Research. She previously served as the commissioner of the Bureau of Labor Statistics. Erica, thanks for being here and walking us through this data. We appreciate it. Thank you very much for having me. Well, we're hearing from you. Jim in Omaha, Nebraska says AI is revolutionizing the field of radiology more than other fields in medicine, but it will never replace the medical expertise of a physician. Coming up, more on AI's potential effect on work and workers as it becomes more sophisticated. That's just ahead.

13:01Stay tuned for the following announcements and messages from our sponsors. This message comes from the UPS Store. The UPS Store is open this holiday to help make sure your small business can stay at its best. During the holiday, customers have access to professional printing, packing and shipping, shredding, mailbox services, notarizing services, and everything else your small business could need. Find nearby locations and keep tackling important tasks at the upsstore.com. Most locations are independently owned. Products, services, pricing, and hours of operation may vary. See Center for details.

13:35This message comes from Squarespace. A beautifully designed Squarespace website is the difference between thinking about photography school and selling your portraits online. It's the difference between giving away cookies for free and running your cookie empire. A beautifully designed Squarespace website is the difference between having a movie-reviewing hobby and a movie-reviewing career. Squarespace. Because no matter what you're into, a website makes it real. Use offer code DIFFERENCE for 50 % off select plans. Terms and conditions apply. This week on Shortwave. What is something found on both true crime TV shows and real-life investigations?

14:12Mm, questionable forensics. And it's almost always accepted by legers as providing an objective scientific truth. Interrogating junk science. That's on Shortwave from NPR. Let's get back into the conversation. We're talking about artificial intelligence and how it's changing how we work and the careers we can have. It's part of our ongoing series pending about uncertainty in modern life. We're hearing from you. Mike writes, I watch people in my office use AI to summarize their emails and craft their responses without the human ever really absorbing the information. It makes for very clunky meetings when it becomes clear how few of the people attending really know what we're meeting about.

14:56We also got this message from one of you. At work, my boss requires us to use AI. As a matter of fact, she has tightened deadlines stating that because we're using AI, we should be more efficient at meeting our tasks and our goals. And a member of the 1A Text Club writes, as a technical writer for software development, we do use AI. It's helpful. The writer becomes a researcher, gathering and verifying information with less focus on the written word. Well, joining us now from New York is Arun Sundarajan, a professor at NYU's Stern School of Business. He studies AI and the future of work. Arun, welcome to the program.

15:35I'm delighted to be here, Jen. And joining us from Washington, D.C. is Liz Wilkie. She's the chief economist at LinkedIn. Liz, welcome to 1A. Hi, Jen. It's a pleasure. Arun, when you look at the AI tools that companies and workers are most commonly using today, for example, generative AI chatbots like Claude or ChatGPT, what kinds of tasks are they good for and what are they not so good for? What I'm seeing is that they are really good for tasks where you need to generate lots of options. You need to ideate. They're not so good at tasks where human expertise is needed in order to verify. And this is not surprising.

16:17I mean, we're in the middle of a revolution of generative AI rather than verificative AI. So in many ways, what this current generation of AI has done is it's dramatically lowered the cost of generating, but it's also raised the cost of verifying. And so, yes, it's really good at simple coding tasks, but the human is still incredibly important at this moment in complex coding tasks and making sure that code that is deployed is secure. it's pretty good at simple customer service. And I think that in the long run, a vast majority of customer service will shift to being AI driven. A lot of simple graphics for sales and marketing.

17:06But as Erica was pointing out, you know, there's the capabilities that AI has are very different from it's deployed, what it's being deployed at. And this is, you know, there's this myth of technological determinism where we think that once the technology is good enough to do things that humans are doing, it will instantly start doing them. But sometimes there's a lag of 5, 10, maybe even 15 years. And so we're sort of definitely in that phase right now. Well, Liz, how are you seeing companies change what they include in job listings when it comes to AI? For instance, looking for specific AI skills.

17:48Yeah, we've seen a lot of evolution over the past several years. It used to be that the fastest growing skills on worker profiles related to AI were much more related to technical building and deployment, things like Kubernetes and Scikit-learn, and very technical terms for how a software engineer or a data engineer might use these tools. And what we've seen over the last few years, especially into 2026, is we've seen those skills translate into more general AI literacy skills have become the fastest growing skills on worker profiles. We've seen this broad diffusion where those AI skills show up outside of technical worker profiles.

18:26Coupled with that, we've seen job postings requesting AI literacy skills, the ability to work with LLMs, the ability to work with co-pilot or Claude or anything like that has also jumped 65 % year on year. So we're seeing a response on both the supply and the demand side for these skills. What kinds of changes are you seeing in the number of job postings for roles that have become more AI-dependent, software engineers, for example? We've seen a broad slowdown across the board. The LinkedIn hiring rate where we track the rate of hires year-on-year has only just become positive for the first time since July 2022.

19:02The rates for things like software engineers and coders are definitely down, but LinkedIn hiring has been decelerating across the board for some time. Arun, we heard earlier that U.S. hiring data so far shows the only really notable hiring changes is among new graduates. Companies might be employing fewer of those workers and instead having AI complete the less complex tasks they often handle. But if that trend continues, it could represent a structural change to our economy. So what could that change mean for the U.S. over the next decade? Yeah, I think we are entering a phase of a few years where we're going to have a lot more unpredictability in the job market.

19:47You know, certainly AI is not showing up as mass layoffs, but it may be showing up as, you know, jobs that never get offered. And as in some ways work that was never a job to begin with. What I mean by that is I think a lot of the impacts of AI on work are being hidden in a significant reduction in freelance work. You know, we don't count freelance work in the same way that we count jobs. And so there are maybe the same number of freelance workers, but they're all getting less work. And that will not show up in the statistics. It doesn't show up so much in the job postings. I expect that a lot more, I mean, paradoxically, a lot more of work in five years is going to be not as employment but as non-employment work.

20:38It's just that the frictions of shifting away from freelance work are a lot lower. And so we're seeing some of AI's impacts there sooner. We got this message from Zach in Pennsylvania who says, I use AI every day as a senior employee. and it's pretty good at replacing an intern or new hire with a lot less friction of having to train someone. This is both good and bad. It makes the work get done faster, but the bench is going to start getting really thin in the coming years if this becomes the standard. And Arun, this is something I've been thinking a lot about. You know, entry-level jobs or internships, they serve as a pipeline into certain professions.

21:17So what does the potential narrowing of that pipeline mean for younger workers and for our economy as a whole? I mean, this is a great question. This is something that really worries me about sort of the sustainability of human capital in our economy, in a sense, because, yes, hiring someone entry level is generally an investment for a company. and it's actually something that larger companies make more of an investment into. It's an investment for the economy as a whole because sometimes you train someone and then you don't get to keep the benefits of that training necessarily because they may go on to a different job.

22:03But the exchange tends to be that you get a bunch of stuff that you don't want to do done. Now, AI is doing a lot of that stuff. Yes, it's only 20 % of companies that are actively using AI. But if you look at an employee-weighted number, most of the biggest companies are using AI very actively. And so we may end up in a world in five years where nobody has made these investments into growing the next bench of human capital. Everybody's thinking, well, we'll just hire them from someone else. And we'll end up with this labor market disequilibrium where we don't have enough human expertise to be able to run the economy.

22:49And this will then lead us to lean excessively on AI, not because it's better, but because we don't have an alternative. Well, I want to go back to our voicemail box. Here's a message we got from Ian in Utah. I decided to study math and more than that, adjacently computer science because I thought it would offer me a better work opportunity as automation becomes more applicable. The irony is that it's kind of done the opposite. And I feel like as someone in the AI and machine learning mathematics space, I'm actually at a significant disadvantage as someone who is, say, a plumber or works in retail or even maybe in finance or something.

23:29Ian, thanks for that message. I mean, Liz, generative AI tools keep changing. ChatGPT in fall of 2022 is not the same product as today's chatbots. So how is the kind of AI tools that workers mention on their profiles, how is that changing over time? Yep, we're seeing an evolution towards the much more practical applications of AI tools. That's both in the engineering case where you're really doing agent development and deployment and coding agents and supervising them, which is a bit more technical than using LLMs to write better emails and marketing copy, et cetera. Both of those are sets of AI skills that we track.

24:09What we do see is there is a lot of evolution. And I think that this illustrates the uncertainty that is on both sides. In a survey that we've done, half of our CEO business leaders say that they are very sure that AI will have a huge impact and that it is a priority, but they do not know how it will change the work and the skills that are required in their organizations to do those jobs. And I think that is really contributing to a sense of uncertainty on the part of workers and on the part of business owners about what do I need to search for? Who do I need to hire for? What we also find is that young workers in particular with AI skills do have better hiring outcomes.

24:49So having AI skills on your profile is associated with 78 % improved hiring outcome. And I think what we will also find is that these senior workers will start to find when they have entry-level workers who do know how to do these AI prompting tasks and who do know how to work with these skills, they will find that they can delegate AI tasks to their interns and to their entry-level workers. And I don't think people have really figured out how to reorganize that work, right? So I think that senior workers may be taking on a lot of tasks that it was previously difficult to explain to someone who lacked that context.

25:26But I think as younger workers start to adopt these skills more and more, we will see some reorganization where more senior workers will still realize that their time is very valuable and they will figure out how to give entry-level workers more tasks that are appropriate for them with these AI tools. Amber emails, I'm not a fan of AI at all and take steps to reduce my usage of it, such as adding keywords and my internet searches to remove AI results. I'm employed in global supply chain distribution and my employer is also taking a firm stance on limiting the use of AI, which I fully support and appreciate.

26:01Auburn, I'm hearing an interesting tension here in that there's a lot of uncertainty both on the side of employees and employers about what AI means long-term for work, for workers, for our economy. And yet there is this sense that for at least some companies, we're going to forge ahead even in the face of that uncertainty. What do you make of that decision on the part of some companies, at least, to push forward even though it's not really clear where this is heading? I think it's an important period of experimentation for companies. And we've seen this with technologies like AI in the past, what economists call general purpose technologies.

26:47Early in their adoption, they tend to speed up certain things, but there's a pretty heavy investment and often the productivity gains don't show up. It takes a few years of experimentation, of figuring out how to do work differently before the gains start to come. And so for a lot of companies, I think they're in that first phase where they haven't yet figured out how to get the true returns from AI. Right. Sort of it's like, you know, if you roll back like over 100 years, there were factories with steam engines. You know, when electricity and the internal combustion engine came along initially, they just sped up the work that they did a little and had a slightly more reliable source of power.

27:40But it took the invention of scientific management, the assembly line, the fact that everybody could have a machine at their own station to really unlock the productivity benefits of electricity. So we're in that same stage with AI. I'm also seeing a lot of uncertainty in larger companies around how many humans they're going to need in the future and in what roles. And so part of the tentativeness that we may be seeing in the job market at the door, not the exit, right? I mean, the entry-level work is where we're seeing the effects. Could be not a long-term trend, but just a company being cautious because they're not quite sure what mix of talent they're going to need in a few years.

28:29I want to read this message we got from Tara in Ohio who says, my company spent millions on an AI program to take over part of our work planning process. It ended up being so inaccurate, even after at least four rounds of verification and algorithm fixing, that the whole thing had to be scrapped three years in a row. At this point, the company has not given up on the project, but it looks like we're heading in that direction. Arun, another really interesting tension is that this technology is at the center of the U.S. economic story right now. So you described a sort of tentativeness on the part of businesses while we're figuring out what this technology can do, what it can't do.

29:08But we're also seeing, you know, startups building AI products, the AI infrastructure buildout that's expected to cost $580 billion this year in the U.S. So help us understand, you know, the overall picture for the U.S., where at one hand there is a lack of certainty, but we've also have this huge reliance on artificial intelligence being successful because of the way it is embedded in our economy. Yeah, I think the macroeconomic story is a really important one at this moment. I mean, AI is certainly driving a lot of the growth in GDP, and that comes sort of bundled with a lot of risks. I think there are a few scenarios here.

29:54One is that there's a massive overinvestment in infrastructure. The AI bubble is going to burst and a lot of companies are going to go out of business. Now, the assets will remain. This happened with internet capacity and fiber optics. This actually happened with the railroads in the 19th century. The companies go out of business, but the assets remain. A different scenario is that AI-driven work and AI-driven products are going to be a much bigger fraction of the economy in the future. While, you know, there was the listener who indicated that, you know, the use of AI in medicine will never eliminate the doctor, that may be true, but it could be that a significant fraction of healthcare spending moves away from today's healthcare companies and moves towards companies that are offering AI solutions, the economy could grow as a consequence.

30:54And so a lot of this spending could end up being worthwhile, even though we are spending like, you know, sort of what is approaching a trillion dollars a year, you know, economic growth and movement away could end up like, you know, sort of producing the return. We have to take a quick break, but when we come back, more from you and our guests. Stay with us.

31:44tasks at the upsstore.com. Most locations are independently owned. Products, services, pricing, and hours of operation may vary. See Center for details. Welcome back to the 1A podcast. Melissa emails, I work as a law clerk summarizing thousand page depositions for my law firm. With the help of Claude and a proprietary AI for lawyers, I was able to take a task that took 40 hours down to 45 minutes over a couple of weeks. really helpful, but still needs careful checking when it's done. We also heard from Rebecca in the 1A Tax Club who said, for 18 months, I used AI at work until I had completely trained it to do my work in 20 % of the time.

32:25Then I taught my coworkers. Then we all got laid off. You know, Liz, we mentioned businesses trying to figure out how to use AI. Workers are clearly learning these tools and figuring out how it fits into their work or not. The Census Bureau found earlier this year that 24 % of U.S. workers who responded had used an AI tool at work every day in the past week. Another 45 % said they'd used it at least one day in the past week. What does this learning how to apply AI to work you've already learned how to do typically look like? What we see is that this is a technology that where it is diffusing and appearing in workers at least as fast, if not much faster than it is appearing in companies.

33:10anyone can get a Claude account or a ChatGPT account or a co-pilot account and start to play around with it. And there are obvious use cases in your day-to-day, responding to emails, summarizing documents, et cetera, then workers can apply their context. There are many, many tasks in our day-to-day jobs that are a little boring, a little frustrating, a little more time consuming. And that is a real use case that I think a lot of workers are experimenting with AI to figure out. how to make themselves more efficient on this sort of work that's not value generative or possibly not that much fun. And then they learn how to use it and then they learn how to do more with it.

33:51And we do see a lot of that. Many, many workers are adding these skills to their profile faster. And one of the fastest ways that these AI skills show up in companies is actually workers picking them up. And so I think what that means is that there are lots of use cases. Whether or not that translates into broader business productivity is an open question. It's one thing to reduce the amount of time an employee spends on email, and that may make their workday a lot nicer, a lot smoother. They may be able to do more value-added work. But whether or not that translates into complete business productivity really is a different question.

34:29I think one of the reasons that workers both see the benefits of this kind of technology in their day-to-day work and also feel uncertain is they don't know the decisions that the business will make about how to organize their work. And as Arun previously mentioned, lots of companies may be themselves very disrupted by AI to the point of going out of business and money will move in the economy to other places. So even if we see net employment and aggregate employment stay the same, I think the uncertainty is that there could be a lot of churn underneath as workers pick up these skills, but then have to port them over to different businesses, right, with different opportunities.

35:09And I think that is contributing to a lot of the uncertainty. The one thing we do know for sure is that workers do recognize that this is valuable and that employers are looking for these kinds of skills and that they are picking them up with or without the support of the businesses that they work for. And they do recognize that these have value in the market. But Arun, when you hear a story like what we heard from Rebecca, someone who trained AI to do her job, taught the co-workers, and then everyone got laid off. You know, someone listening to this conversation might think, okay, on one hand, I need these skills.

35:45It makes me more marketable in the economy, but I also want to protect my place in the economy. So do you have any practical advice for how people can sort of thread that needle? Yeah, it's a really difficult balancing act. I think, I mean, to Liz's earlier point, being an AI force multiplier, being someone who is actually able to use AI to dramatically up your productivity, but in a way that produces good work, you know, so being able to sort of verify, learning how to verify, like, you know, what the AI produces, that's going to be really important in the future. But you raise a really important point, Jen, which is that we seem to be ceding ownership of our human capital to some extent to AI systems.

36:38This is happening tacitly to some extent, you know, in the kinds of examples that like the listener brought up. It's happening explicitly where people are being paid hundreds of dollars, like, you know, doctors, lawyers, scientists, mathematicians, you know, people with MBAs to actually generate expertise that trains, you know, the next generation of frontier models. And this is good in the short run for the people who are getting paid. But in the long run, you begin to wonder whether people are going to invest in human capital in the future, if there isn't a way to sort of like, you know, get value from it.

37:20And so it's the kind of thing where, you know, I think society has to come up with a way to balance. It's a very hard decision for an individual. I would say, constantly keep trying to position yourself and to pick up the skills to be able to work productively with AI, to be able to produce more from AI. And, you know, through learning by doing to be able to produce stuff that is reliable. The example that the listener came up with, like 45 minutes versus 40 hours, that works if you can actually stand behind the work that is generated in 45 minutes. And if you're one of those people, then you're going to have a good future in the workplace.

38:09Liz, I'm also thinking about skills that AI can't take over, interpersonal skills, a decision-making ability that comes from years of experience working in a field. How can workers also optimize that part of their experience when they're either seeking a job or trying to hold on to a job? I think there's no current substitute for the two types of skills you mentioned, which are what you might think of as softer skills, communication, leadership, coordination, collaboration, as well as the judgment and the domain expertise that allows you to make a decision in context with confidence. What we do see a lot of on platform is individuals showcasing both sides of that coin.

39:01They will talk about a project that they have done with the use of AI. They will link it to a problem that it solves in context with relevance, and they will talk about how they worked with other people in order to bring that all the way to fruition. There is no current curriculum for proving that you can combine AI with business context, with soft skills. And until we have some standard way to do that, I think we will continue to see workers trying to showcase that in a variety of formats. And we do see it happening on the platform. But I think that the value of AI as a technology rests entirely on the ability of workers and process to turn that technology as an input into value delivered for a client or a customer, right, that they're willing to pay for.

39:52So I think that is at the crux. We have this new technology. We know how to create value in an old world without this technology. and workers now need to figure out how to create the same or new value with this technology and to demonstrate that they can combine those soft skills, that judgment and expertise that comes with just knowledge and context with this technology embedded. We've gotten several respondents from people who have a very negative view of AI's use in the workplace. Others wrote in with positive use cases for AI. Arun, what would you say to someone who says, I don't believe AI is good for my work or my company, either because I'm against it on principle, or I just don't see it providing much value to the work itself?

40:39You know, they may well be right that in the short run, it may not seem to them that AI is producing value and AI may actually be lowering productivity in the short run. But the thing that they have to keep in mind is that over time, AI's adoption in the workplace is going to deepen. We're 20 or 30 years into AI adoption already. It's just like, you know, it's been four years since ChatGPT, but AI has been around for a while. And, you know, hiding your head in the sand, in a sense, by saying that I don't like AI. Resisting AI is probably not a good long-term strategy. You know, instead embracing AI and trying to understand, as Liz pointed out, you know, how to work with AI and how to demonstrate that you are one of those people who knows how to extract value in an AI-powered world, that's a much better recipe for future labor market success.

41:42being entrepreneurial and being flexible and, you know, building the muscle of being able to do things on your own. Even if you don't start your own company, that's going to be a really important skill in the future as well, because a lot of what we do in five years as human beings is going to be different from what we do today. And the people who know how to adapt to the new needs that are going to be demanded from humans as AI starts to take over some of the things that humans used to do. Those are also going to be people who are successful. So I'm sympathetic to people who feel like threatened by AI.

42:20It's certainly a threatening time, but you have to sort of embrace the reality that it is going to be omnipresent in the future and you have to learn how to work with it. Arun, I just want to circle back to this idea that AI adoption is inevitable because I think there are people who just don't accept that premise on its face. They say AI adoption is a choice. Businesses choose to adopt it. Individuals choose to use it. States and the federal government choose to invest in it in the form of data centers and other infrastructure. It's not inevitable. It's just choice. And so I think part of the struggle people have and part of this feeling of uncertainty is that they're feeling pushed towards something that is not the only option available to us.

43:17And I'm wondering how you're making sense of that. I mean, that's, again, another great point. Of course, it is a choice. you know in many ways the technology enabled future is always a choice you know I mean there's we don't make predictions about the future in some ways we are making choices that end up shaping this future but you know if you go back to the early 1980s when personal computers were just emerging in the workplace and a lot of people didn't like them this happened in the 1980s and some people Like, I never want to use a personal computer. I don't like this computer thing. I like talking to people.

43:59I like, you know, talking to my clients on the telephone. You know, and people like me at that time would have said, well, you know, you should embrace computers because computer use for knowledge work is inevitable. And that's the stage we're at with AI today. of course it's a choice but you know one can confidently predict that it's a choice that pretty much any company that deals with information or deals with knowledge as part of their work is going to embrace and so it doesn't matter if you're a doctor it doesn't matter if you're a lawyer doesn't matter if you're doing supply chain whatever profession you're in there is going to be a significant component of what you do that the people who design your work for you are going to choose to delegate to AI.

44:50And so you want to be one of the people who is comfortable with that choice and who knows how to amplify the benefits, who knows how to sort of, you know, allow your company to get greater returns from AI. We got this from Mark in Burlington, who listens on WUNC. As an IT engineer at a Fortune 500 company, AI and unreasonable management expectations are headed to an industry-wide burnout event. The hopes for AI were to reclaim time for an already overwhelmed IT team. The reality is that reclaim time is now spent producing more and being judged by invalid performance metrics to keep your job. The work to prove you add value to the AI platform is never-ending and self-defeating.

45:36And I'd love to hear both of your responses to that. We have about 30 seconds left that for some workers at least, this is not adding to more productivity for them or more freedom in their work. It's leading to burnout. Liz? I think that that's a really understandable reaction. I think Arun has mentioned this as well, but we are in a place where there's a lot more experimentation and unknowns than there is clarity about how exactly AI can be integrated in a way that's productive and effective for companies and for workers. I think that many, many workers feel sort of caught in the in-between. What we also know is that as many workers feel sort of negative about the situation and how to work through it, we are seeing that confidence in AI is increasing.

46:24And as they take on more skills, that improves confidence. Arun, really briefly, just a couple of sentences. I think it has to do with incentives that companies give you. A lot of companies are saying use AI without thinking about how the AI is going to be used. that leads exactly to the kind of burnout that the listener mentioned. It's a phase of experimentation. It's going to get better. Well, that's Arun Sundarajan. He's a professor at NYU's Stern School of Business. He studies AI and the future of work. Also with us, Liz Wilkie. She's the chief economist for LinkedIn. Liz, Arun, thanks for speaking with us.

46:58Remember, this is part of our ongoing series about uncertainty called Pending. You can listen to past shows in the series about faith, family estrangement, quitting, and more. Head over to the website v1a.org and check the series tab. Today's producer was Michael Falera. This program comes to you from WAMU, part of American University in Washington, distributed by NPR. I'm Jen White. Thanks for listening. And we'll talk again tomorrow. This is 1A.

47:43Stay tuned for the following announcements and messages from our sponsors. This message comes from Mint Mobile. Unwrap huge savings this season. All plans, even unlimited, are$15 for new customers. Go to mintmobile.com slash switch. Upfront payment of$45 for three-month,$90 for six-month, or$180 for a 12-month plan required,$15 per month equivalent. Taxes and fees extra. Initial plan term only. Greater than 50 gigabytes may slow when network is busy. Capable device required. Availability, speed, and coverage varies. See MintMobile.com. The number one issue in America right now is the economy. Whether it's the high price of gas, a potential AI bubble, or the unsteady job market.

48:26At NPR's The Indicator, we've got you covered. Understand the economy in 10 minutes or less each day. Listen to The Indicator from Planet Money on the NPR app or wherever you listen to podcasts.

From the publisher
“Today, the convergence of automation, robotics and A-I is rapidly transforming the very structure of work. It is said that this will bring great improvements for everyone. In reality, however, the “new ways” of working are not necessarily better,… …it is necessary to design systems that are centered on the human person and not solely on performance.” 

That critique of artificial intelligence is one you might expect to find in an op-ed or an online forum. It’s actually from Pope Leo the XIV, writing recently about protecting humanity in the AI era.

The risks that generative artificial intelligence pose to human work are a common source of worry. A summer survey from Pew showed a majority of respondents – 52 percent – are more concerned than they are excited about the technology. And for the first time in five years of asking the question, a majority of people surveyed under 30 years old said they were more concerned than excited.

How much of this worry about work and AI is based in reality? How are employees changing their work to include AI?

Find more of our programs online. Listen to 1A sponsor-free by signing up for 1A+ at plus.npr.org/the1a.

See pcm.adswizz.com for information about our collection and use of personal data for sponsorship and to manage your podcast sponsorship preferences.

NPR Privacy Policy

More from 1A

All episodes