The next president’s AI problem

I'm NFL linebacker TJ Watt and this is my personal best. YPB by Abercrombie is the active wear I'm always wearing. That's why I reached out to co-design their latest drop. I worked with designers to create high-performance activewear that holds up to my toughest workouts. Shop YPB by Abercrombie in-store, online and in the app. Because your personal best is greater than anything.

Hey, welcome back to Politico Tech. I'm your host, Stephen Overly. And on this show, I break down tech politics and policy with the people shaping our digital future. With all of the headlines about AI leading to job losses, there's been an urgent question on my mind about how to prepare workers for the major changes ahead. Now, we can debate how high unemployment will be.

and whether that will be offset by the creation of new jobs. But the disruption to the economy, and to individual people, is already happening. And my guest today thinks it could affect millions of workers in just the next two years. Bharat Ramamurti was the deputy director of the National Economic Council under President Biden. And in a recent opinion piece, he called AI, quote, the biggest economic issue no one is talking about.

On the show today, Bharat and I delve into the options he thinks policymakers have in front of them, and why he thinks economic anxiety driven by AI will dominate the next election cycle. Here's our conversation. Bharat, welcome to Politico Tech. Thanks for having me. You know, I have to say this interview for me is a long time coming because I actually set out over a year ago to talk about.

how to support workers who are going to be displaced by AI. And I couldn't find any policymakers who were like seriously grappling with that issue. Why did you start digging into this? I think it started from the same place, which is AI is the talk of the town. Everyone has a view on how much productivity it will cause, how much growth it will cause. And people didn't seem to be...

grappling with the fact that the way that that productivity increase is supposed to happen is by changing the way that people work and the types of jobs that people have. And as I looked around at the types of proposals that policymakers had put together to address this sea change in the way that our economy operates, I was really surprised to find that no major national politician was trying to address it.

And so part of the reason that I wanted to write something on this topic is just to urge policymakers to start putting their ideas out there. We need to have a big national conversation on this topic. And we're, I think, already behind the curve, given the pace at which AI adoption can happen.

Well, and there's this debate, and it's still a debate, I think, about whether AI is going to replace jobs, whether it will simply change jobs and how we work. You know, I think we already see signs that both of those are true to some degree. How can you be sure that AI is responsible for job losses and not other economic factors?

Look, I think the data is going to be a little bit difficult to parse, especially at the beginning. It is hard to say definitively that, for example, the recent slowdown in hiring of college graduates with certain types of degrees. is purely because of AI adoption. But I think that there's two important things. Number one, if we are to believe that AI is going to cause a big increase in productivity and growth in the United States,

That necessarily implies that it's going to cause big changes in the job market. The other thing is that there is good evidence that when we enter an economic downturn, Companies use that downturn as a reason to switch to automation. The data that I've seen says that 90% of job losses from automation historically have happened during the first year of a recession.

So if you're looking at maybe 10, 12, 15 million projected job losses from AI, which is what a lot of the consensus is, you could see a big chunk of that in the next year or two. If in fact, we do go into a recession. The third point is just companies are going to have an incentive if they're cutting back jobs, even if they're not adopting AI, really.

to be out there saying this is because of AI. Because that's a better story for their investors to be saying, it's not because we're trying to tighten our belt, it's because we're using this newfangled technology and it's because we're a cutting-edge company. And regardless of what the data says, at the end of the day, a lot of people are going to be hearing.

AI is causing these people to lose their jobs and AI is threatening my job. And so as a policymaker who wants to be responsive to public concern, it seems to me that policymakers should be out there getting ahead of this and saying, here's what our plan is for addressing. potential job losses from AI and protecting the interests of American workers.

The impact of AI on job creation is something that I think is often overlooked. You know, I started my career in journalism writing about startups, and it used to be a badge of honor, right, when a company was scaling and hiring 200 people and look at all this growth. And now when you see startups, the badge of honor is, look how much we're scaling without hiring people. Look how much we're using AI to grow without needing bodies and payroll. Yeah, exactly right.

I think at a minimum, I feel safe saying that AI is going to change the composition of jobs in the United States. But it's also likely to shuffle around the mix of jobs in the United States. So you could say, okay, we're going to lose. Several million jobs in fields like computer science and coding, accounting, finance, call center operators, customer service, but it's also going to create new jobs in other areas.

That's great for the economy overall if people are shifting from less productive jobs to more productive jobs. It's not great for the people who are affected because you can't snap your fingers and say, I'm a 50-year-old accountant. This is what I do.

But now there's all these new jobs being an electrician, building data centers. So all I got to do is learn how to be an electrician in my 50s. That's not how it works. And as we've seen in the past... for example, with globalization and NAFTA and the effect on American manufacturing in the 2000s, that can be a profoundly disturbing experience for a lot of workers, and it can have big political ramifications.

Well, that's exactly where I was going with this, because in addition to covering tech, I used to cover trade. So this is all very much my domain and my interest area. There have been attempts by the government in the past to support displaced workers. And obviously, you're referring to trade. nerds will know as trade adjustment assistants write programs in the 90s to help workers in manufacturing whose jobs were outsourced because of globalization. That really...

was inadequate in many ways or didn't deliver as promised. I think even more recently about payment protection loans for workers displaced by COVID and fraud that happened to that program. And so... It raises the question for me of like, is there anything that can work? It's a good question. I agree with you that what I think of as retroactive, retrospective relief, you know, in other words.

You're a person who lost their job because of AI, and so the government's going to come in and give you some kind of compensation or help. We'll give you some money to get through this period, or we'll give you access to retraining programs or upskilling programs so that you can get on a new career path. The track record of those types of programs, as you've said, is very poor.

We don't fund them well enough. When we try to do the reskilling, we find that workers are often behind the curve relative to other people who've been working in those fields for a long time. What I think would be useful is

doing something to empower workers on the front end. And here, I think there's an important contrast with Europe. So in Europe, they have a much different type of relationship between labor and management. In many countries in Europe, you have workers on boards of directors at corporations. You have much higher unionization rates.

You have what are called works councils at the firm, which negotiate directly with management. We don't really have that infrastructure in the United States. The unionization rate in the private sector is only 6%. We need to figure out a way, in my opinion, of creating some kind of bespoke tool

that will allow workers to articulate and effectuate what they want out of this type of transition. I think that that will not only lead to better financial and economic outcomes for workers, I also think that there's an important autonomy element here. which is that people who've been preparing for life in a certain field or have been working in a certain field for a long time don't want to feel like the rug's being pulled out from underneath them without having any say in that.

Well, the counterpoint that I'm sure you can anticipate, because we hear it often, whenever you invoke Europe, right, is Europe wishes they had the... innovation, corporate culture of the United States. They wish they had a Silicon Valley or sort of the technological powerhouse that is the U.S. I mean, are these things at odds, empowering workers to have a voice in AI?

and the culture of business in the US? Look, I think there's a number of factors that go into these disparities between the US and Europe. My view is that if you think that Europe has gone too far, in terms of empowering workers, and that has led to less innovation and less growth. I think that's a questionable argument, but let's say you believe that. I think it's clear that the U.S. has been too far on the other side.

of the pendulum, where our focus on growth and innovation has led to less than ideal outcomes for workers. To me, one of the key findings is that over the last 25 years, automation has led to a declining share of national income going to labor. To me, the challenge for policymakers is with this big technological revolution happening. What is our pathway for actually using this to raise the share of national income that goes to labor?

in the form of better wages and better benefits? How do we defy the trend of the last 25 years? And yes, I think it makes sense to at least consider some of these worker empowerment efforts in order to get that outcome, because the alternative, frankly, is... to have more and more of the money go to capital and then to try to redistribute it through tax policy. There's all sorts of issues with that. It's not like that's politically popular with a large set of folks.

When we do it, we tend not to do it in a way that actually matches the scope of the problem that we have and the amount of money that we would need to transfer back to the working class. And frankly, I think a lot of the transfers that we have are very narrowly concentrated on the very poor.

And we don't get to working class, middle class people that are actually also going to need some help. For some workers, it may not be the case that they want additional $1,000 or $2,000 a year. They want to keep their job. or they want to keep their job, but they want to work less. There's different outcomes that we want to effectuate. The way we can do that is by empowering workers. This episode is brought to you by State Farm. Checking off the boxes on your to-do list is a great feeling.

And when it comes to checking off coverage, a State Farm agent can help you choose an option that's right for you. Whether you prefer talking in person, on the phone, or using the award-winning app, It's nice knowing you have help finding coverage that best fits your needs. Like a good neighbor, State Farm is there.

That idea of increasing corporate taxes, because it's been talked about, right? This idea of if you employ AI instead of a human, should companies still pay some kind of tax on that? It sounds like you're skeptical of that working. I mean, again, is it a good idea to specifically tax AI adoption? I'm a little skeptical of that in the sense that if we think that AI adoption can be good for economic growth and productivity, which means in theory more for all of us.

We don't want to discourage that necessarily by putting a tax on it. Second of all, if you think that AI adoption is going to be very widespread and basically every type of company is going to use it one way or another. then we should just raise corporate taxes generally. There's various permutations of that. But, you know, we've seen the corporate tax rate already go down from 35% to 21% because of the Trump tax cuts.

And so do I think that there's room to raise the corporate tax rate back up? Sure. I think that that's a viable option. We tried to do that in the Biden administration, and we ran into objections from Senator Sinema and Senator Manchin. Even with Democrats who are broadly on board with raising corporate taxes, there's going to be some amount of pushback on that.

nervous about the idea of having our whole response be conditional on, are we going to be able to raise corporate taxes enough to actually compensate people for the changes that they're about to experience? My preference would be to try to Adopt a set of policies that allows that labor share of income to go up over time and not just try to address it on the back end through taxes and transfer policy. It's funny you say that just because I obviously read a lot of tech news and I saw.

Two back-to-back headlines recently. One was new tech graduates. can only find jobs at Chipotle. And then the next headline was AI is going to create more billionaires than we've ever had before, which just kind of speaks to the bifurcation of AI and its economic impact. You mentioned raising corporate taxes, not being politically popular with.

some and Republicans, as I think who you're referring to there. It is interesting, you know, just recently, we've seen the Trump administration reportedly make this deal with microchip makers, NVIDIA and AMD to take 15% of their sales. to China. I know this very much used to be kind of your wheelhouse when you were in the Biden administration. I would just be curious what you make of that approach. Do you view it as like a new corporate tax, a national security risk, something else?

I'm all for getting more revenue from corporations. I think we've gone way too far in reducing the amount of revenue we bring in from corporations. are basically saying we are going to jeopardize both economic leadership and potentially national security by allowing the sale of these important AI chips to China. What we're going to get in return for those sacrifices is maybe $1 billion or $2 billion in revenue.

that doesn't seem worth it to me. And I think it's a question of where do we draw the line then? Are we going to be okay with selling... military equipment and jets to some of our rivals internationally, as long as the companies that are making those sales are going to give us 10% or 15% of their revenue. I just think it's a dangerous path to go on. And so I would be in favor of having higher corporate taxes.

overall, and then having some sort of export regime that makes sense given our national security and economic security interests. I don't think that what the Trump administration is doing here accomplishes that. Second of all, I'm pretty sure it's unconstitutional. Article one of the Constitution says that you can't have a tax on exports. It says that extraordinarily clearly. I don't really know how you interpret this other than a tax on exports.

Literally, the government taking a percentage of sales from a domestic company abroad. That seems like a tax on exports. I don't know who, what entity would sue in this situation to actually get that kind of determination, but it seems to me very clearly to be unconstitutional as well.

I'm curious, now that we're six, seven months out of the Biden administration, I know you left before that, but... When you look back on the Biden administration, is there more you think that could have been done to get ahead of some of these issues we're talking about, particularly around AI, its economic impact and the way the public is sort of.

receiving this new technology. Yeah, I wish we would have gotten ahead of it. That said, with this kind of technology, and we've already seen that the path of its progress is not linear, necessarily, and a little bit hard to project. It did feel like we were in the early stages there in terms of its impact on the labor market back in 2022 and 2023. We now have a lot more information, I think, about what these models could look like.

what the potential job impact could be. But I think a lot of the things that we tried to pursue in the Biden administration weren't AI specific, but would have been helpful. in this regard. We wanted to raise corporate taxes. That would have helped bring in more revenue that we could use to then help workers who've been affected. We took a number of steps to try to raise private sector unionization rates.

I think the challenge that we have as a country, frankly, we have seen an enormous amount of growth and innovation over the last couple of decades, but it is more and more concentrated among those at the very top. And my warning to those people would be, if you think it's important to have broad AI adoption and to continue to have this kind of pro-innovation approach in the United States.

I think it is important to figure out ways of making sure that working people are getting their fair share of it. Because I think the alternative, and you're starting to see it a little bit at the state level, is just outright bans on AI adoption. You see that in trucking, for example. There's bills in Indiana and other states that say no autonomous trucks. Every truck has to have a driver in it. As this develops over time, could you see more and more of those sort of targeted...

AI bans in certain areas. No AI for accountants, no AI for investment bankers, no AI to replace call center workers. There's going to be political constituencies for those types of bills. The way that folks in the industry and in tech can get ahead of that is by saying, we actually have a better way. We want AI adoption because we can all benefit from it, and here's how that's going to happen.

And so if they're not offering those alternatives, I think the alternative course that we're looking at is going to be a lot of prohibitions entirely. This is a Politico podcast, obviously. So there was a line in your recent commentary that stood out to me. You said job losses and disruption from AI could be the biggest economic issue in the 2028 presidential election. Now, you're an economic messenger. That's what you were doing at the Biden White House.

What is the message Democrats need in 2026 and 2028 around this issue in order to be successful in a way that, frankly, they weren't in 2024? I think they absolutely have to say. We stand on the side of American workers. We're not anti-AI adoption, but we are going to make darn sure that this time around, unlike what happened with trade, unlike what might have happened with previous rounds of automation.

we are going to make sure that American workers come out ahead and benefit from these types of changes. And so far, I haven't really seen any of the leading politicians be that clear. And certainly nobody has come out with a plan for how you can actually make that happen. But I think that millions and millions of Americans over the next year or two are either going to see their jobs change dramatically because of AI. They're going to see their jobs disappear.

because of AI, or they're going to hear about other people losing their jobs because of AI. And that's why I think that you combine that with a potential recession in the next year or two. I think that you can go into 2028 where a very broad swath of people in the United States, their top concern is what's going to happen to my job. The other thing that I think is interesting about AI is that white collar workers are the most

at risk of having their jobs change or their jobs disappear. And of course, previous rounds of automation have largely affected blue collar workers. And so you actually could have this moment of unique solidarity between white collar and blue collar workers, where for the first time, people who work in accounting are saying, you know what, I understand what you're saying.

you know, at that auto plant in Michigan about why it's important that we have more power in the workplace. And I think it's incumbent on Democrats as a Democrat. to seize that and get ahead of it and be identified as the party that wants to stand on the side of working people in this moment. Listen, Bharat, appreciate you being here on Politico Tech. Thanks so much.

That's all for this week's Politico Tech. The show will be off next week, but I'll be back in your ears on Thursday, September 4th. Until then, be sure to subscribe and recommend the show to a friend or colleague. And while you're at it... Subscribe to our newsletters, Digital Future Daily and Morning Tech. Norma Malichal and Cara Tabor produced today's episode. I'm Stephen Overly. See you back here on September 4th.