AI is transforming the labor market — but what happens to the federal government’s revenue when jobs disappear? In this interview, RAND’s Carter Price, breaks down a new report he co-authored with colleague Akshaya Suresh examining how AI-driven labor disruptions could threaten the roughly two-thirds of federal revenue that comes from taxes on wages and payroll.
September Only: The Daily Briefing That Cuts Through the Noise on Your Money, Your Health & Your Retirement
For the next 30 days only — all of September — The Morning Pulse daily newsletter is open to new subscribers. Save 20% with code BRN20.
The Morning Pulse is the Broadcast Retirement Network’s daily newsletter, delivering the stories that truly shape your money, your health, and your future:
✅ BRN programming, seven days a week ✅ The news on aging, finance, lifestyle, privacy, retirement & wellness ✅ Curated by an expert with 32 years of experience — just the facts, written by a human, without the use of AI ✅ No ads. No sales pitches. ✅ Your subscription keeps BRN’s daily program on the air
See today’s issue here: https://mailchi.mp/themorningpulse/the-morning-pulse-the-retirement-news-that-matters5-826268
Normally just $4/month. This September only, take 20% off with code BRN20 — but the window closes when the month does.
Subscribe today: https://buy.stripe.com/4gw9CS5NI8cibVm7su
Transcript:
Jeffrey Snyder, Broadcast Retirement Network
Well, Carter, it’s so great to see you. Thanks for joining us on the program this morning.
Carter Price, Rand School of Public Policy
Thanks so much for having me.
Jeffrey Snyder, Broadcast Retirement Network
And I have to say, I really like this angle, not because it’s an angle, but we have been talking about artificial intelligence. It is being talked about in so many different industries, so many different avenues. Let’s talk about, and certainly a lot of work has been done to assess the impact to the labor market and what that means for you, me, and other people that work day to day.
But you and the team at RAND have actually taken a look and assessed what the impact could be to federal revenue. Tell us about the report.
Carter Price, Rand School of Public Policy
Yeah, so my colleague, Akshaya Suresh, and I looked at, how would disruptions to the labor market caused by AI impact federal revenue? And so it turns out that about 2 3rds of federal revenue come from the wages and taxes on the wages and salaries that people make, or on payroll taxes that come out before you even get your check. And because of that, any disruptions to the labor market caused by AI or anything else would have an outsized effect on the federal budget.
Jeffrey Snyder, Broadcast Retirement Network
Yeah, I’m thinking about, and you can obviously educate me. You mentioned payroll taxes. First of all, I’m thinking about Social Security and some of the other programs like Medicare, et cetera, that people are dependent on, that’s number one.
Second, I think about defense spending, which obviously that’s a core tenant of the federal government. And then there are probably a multitude of other areas. This is not just a disruption to people, it’s also a disruption, I would think, to the institutions that support us, both at the state and also in the federal level, which the report indicates.
Carter Price, Rand School of Public Policy
Yeah, so a lot of people have written about these jobs are more exposed and these people are gonna be replaced by AI and that. And then we took the next logical step was, okay, so if that happens, then what else happens? And so if you do start payroll taxes fund, like you mentioned, Social Security and Medicare, and unemployment insurance and a few other things.
And so disruptions to that, we’re already, we’re $40 trillion in debt as of a few weeks ago. And this could be a real problem to our fiscal situation because we are so heavily, our federal revenue is so heavily dependent on taxes on labor.
Jeffrey Snyder, Broadcast Retirement Network
Yeah, well, I mean, I just think about some of the media outlets that have been covering Social Security, that the trust fund, I’m just thinking about that from a retirement perspective, it’s already scheduled to be depleted by 2032. This would just be a cataclysmic decline that would probably result in a lot of benefit cuts if you play it through. How likely, like taking a step back, I know just anecdotally that people are gonna perhaps maybe lose a type of job or a type of job will be removed because artificial intelligence will be able to automate it or replace it, but then won’t those people have an opportunity to go into another vocation and another work where there would be taxes?
Or do those things, maybe those things don’t happen concurrently?
Carter Price, Rand School of Public Policy
Yeah, so we did look at that and we looked at essentially four different scenarios based on whether or not people are able to find new jobs. So one of the things that economists will tell you is that there have been a lot of labor disruptions in the past. There was the industrial revolution and there have been computers, all kinds of automation.
And people have always, people are still employed. People still have jobs. Now, a lot of these jobs were, social media influencer was not a job 10, 20 years ago.
It’s a job now. Podcaster might’ve been a job 10 years ago, definitely not 20 years ago. And so those kind of jobs, new jobs were created as a horse buggy maker.
Still probably, people are still probably making horse buggies, but it’s not a booming industry as it once was. So there’s always that kind of churn. And the question is, will this time be different?
A lot of the techno optimists seem to think so. Economists are a little more skeptical. And we looked at both in either case.
The same thing with another important factor is who owns the AI and how is it priced? So if it’s priced monopolistically because there are one or two companies that are in charge and they make a lot of profit, then that’s gonna have some implications for corporate profits go up. We do tax corporate profits.
And so that would be in some sense good for federal revenue. On the other hand, if it’s freely available, or available for some nominal cost, because there are 10 different companies or 100 different companies providing AI services, then the cost would go down, much lower profits, much less revenue from that angle. So we looked at all these cases to just sort of explore what the outcomes might be.
Jeffrey Snyder, Broadcast Retirement Network
Yeah, and I would imagine that, I know that you have done a lot of writing. Your work appears in many major periodicals and you also testify in front of Congress members in different committees. I would imagine, I don’t know this for a fact, but I have to think AI has to be top of mind for policymakers.
And look, Congress’s job is to tax. So I would think that they’re probably thinking about this in some way, don’t you think, in terms of, hey, we’ve got all these things, we’ve got a fund, we’ve got this deficit of $40 trillion. How do we deal with this?
And oh, by the way, we also wanna get reelected every two years, right? So are they contemplating all this based on your estimation?
Carter Price, Rand School of Public Policy
I think it’s starting to be an issue that’s coming to mind. People are starting to pay attention to it. We haven’t seen layoffs because of AI.
There have been companies that said, yeah, we’re laying people off and yeah, it’s because of AI. But the evidence isn’t there. Like you can say you’re laying people off for any reason and certainly it looks a lot better if yeah, we’re laying people off because of AI as opposed to our business isn’t doing well.
So people are saying those things, but it’s not in the evidence. We’re not seeing that in the data. We’re not seeing that the jobs that are most exposed to AI where you would expect to see people being replaced by AI, we’re not seeing that happen.
And companies that say that they’re, oh yeah, we laid off a lot of people because of AI, a lot of them are hiring people back because it turned out that at this point in time, AI isn’t able to do all of the things that people say it can do in at least reliably enough that it can actually replace workers.
Jeffrey Snyder, Broadcast Retirement Network
Yeah, it doesn’t really have a personality, at least not yet. And it’s not, I guess, sentient. But my understanding is these large language models, like every few months, their capacity doubles or triples or whatever the number is.
So it’s growing exponentially. Would this report then, would the deficit and some of the challenges be exponential from kind of, when you’re an economist, you kind of look at things at a point in time, right? And maybe you should do some pro forma things.
But wouldn’t, if AI is growing exponentially, wouldn’t these challenges that you articulate, you and your colleague are articulating, they’d be even greater?
Carter Price, Rand School of Public Policy
So, yeah, we’re not seeing it yet. We could easily start seeing it very soon. And that’s why we did this analysis sort of prospectively.
So this isn’t a problem today. Now, you might say that the federal debt is a problem today. Yeah, it’s a problem, yeah.
And the deficit. But the AI displacing federal revenue is not a problem today. And, oh, go ahead, I’m sorry, finish your thought.
Jeffrey Snyder, Broadcast Retirement Network
I’m so sorry for interrupting.
Carter Price, Rand School of Public Policy
No, no, no, yeah. But that displacement could be a problem soon.
Jeffrey Snyder, Broadcast Retirement Network
I was just gonna ask you about kind of the global implications. And I know it’s probably not baked into your report, but China is our biggest competitor economically. They’re also one of our biggest competitors, if not our biggest competitor on artificial intelligence.
How does that play into your thinking and the team’s thinking regarding this report?
Carter Price, Rand School of Public Policy
Yeah, so that’s not something that we’ve studied. We do have, there are other organizations that have started doing some preliminary work on that. So the Windfall Trust has done some work in that space.
And what makes it, every country is a little different in terms of their mix of workers. So their mix of industries. So some industries are likely to be more exposed to being replaced by AI, whereas others are likely to be more robust or might even be augmented by AI.
And so every country has its own mix of industries and workers. And so it’ll hit every country a little differently. China has, they make chips, they make models, they make sort of all of those things.
We make many of those things. Taiwan makes chips, South Korea makes memory. There’s a whole sort of ecosystem around this.
And how that affects things is sort of, we don’t quite know yet, but the companies that you would expect, the companies that make memory or chips, as they become more profitable, those profits will tax. And so they’ll be, those countries will be a little bit more resilient to shocks, whereas a country that doesn’t do those things and has a lot of knowledge workers that are displaced by AI could be particularly vulnerable. And so, yeah.
Jeffrey Snyder, Broadcast Retirement Network
Again, I apologize for interrupting. I didn’t mean to do that. I was gonna ask you, so do you think that this, in the scheme of things, in terms of issues, I mean, there’s always issues that go into elections.
There’s always issues that kind of drive the American consciousness. This has to be one of the, close to the top. National defense, obviously, some of the other economic concerns, but this has to be near the top.
Policy, regulation, taxation, along with the debt, that has to be close to the top. I’m not saying it is. I’m just interested in your opinion.
Carter Price, Rand School of Public Policy
Well, certainly, like, I think this is something that because of the consequences, that this is worth paying attention to. We don’t know, you know, will CHAT GPT-6 come out tomorrow and replace everybody’s jobs, or will it come out and it’s a dud? You know, we don’t know yet.
And so, that said, there are some things like the national debt that become much more of a problem if this revenue, if there is this shock to federal revenue. And so, things that we can do to prepare for this would include taking steps to reduce the federal debt, or at least get it under control. Because right now, this isn’t a problem.
If it’s a problem in six months, if it’s a problem in two years, if it’s a problem in five years, and we haven’t tackled the national debt or gotten our deficit under control, then it’s gonna make it that much harder to deal with, because we won’t have, we won’t be able to borrow trillions of dollars to retrain people, to provide unemployment insurance, to provide benefits to people if there’s this shock. And so, things that we can do to prepare now for this possible future, and it’s not like we’re gonna feel bad, like, oh no, we paid the debt down and it turned out five years from now, AI didn’t replace everybody’s jobs. We’ll be better off anyway.
So, this is sort of a no regret strategy, or a no regret thing to do.
Jeffrey Snyder, Broadcast Retirement Network
Yeah. Well, I think a lot of people are looking at this, if you’re the average individual American, you don’t really have a say in terms of how fast AI is gonna, what it’s gonna do, and how it’s going to impact your life in particular. But you do get an opportunity to vote in the polling place, and you can also pick up the phone, or actually, I don’t even know if people call their congressman anymore, but they actually may email, or actually they may probably go on Twitter, and, or X, and say something to them.
Carter, we’re gonna have to leave you there. Great report, great analysis, great angle. We look forward to having you back on the program again very soon, sir.
Carter Price, Rand School of Public Policy
Thank you.
