A (2:37)

The combination of those two things, the.

C (2:40)

Possibility of a crisis or displacement in which AI is seen as central, and the lack of clear party divisions means that the new technology may be in a unique position to reshape politics. And that is what we're going to talk about today.

A (2:54)

With me to do it is David.

C (2:55)

Beiler, public opinion researcher and VP at National Research Group. He formerly worked as a data journalist at the Washington Post and joined me on the forebearer to this podcast with David. Welcome to GD Politics for the first time.

B (3:07)

It's great to be here. I'm so happy to be talking to you today.

C (3:10)

I am happy, too.

A (3:11)

So, time permitting, we're also going to get into some questions about the use of AI in polling and a good data, bad data or not data question that we got from a listener. But let's start with the politics of AI. So I reached out to you at the beginning of September and was basically.

C (3:25)

Like, I think you're really smart.

A (3:27)

And I've got a new podcast that I want you to come on. What's on your mind when it comes to politics. And you sent me a stream of consciousness in response about how AI could reorient politics. And my reaction was, this is fucking fascinating.

C (3:41)

But then we quickly got into election.

A (3:43)

Mode and we didn't manage to create a show out of it. So today is our day. This is our time. Let's start with where the political divisions over AI stand today. Maybe the GOP versus Democrats.

B (3:58)

So one useful way to think about this is almost in terms of both parties, like a pyramid, right? The base of each party is, is it's voters. That's, you know, the fundamental thing that you need in order to have a party. And as you go up, sort of in levels of activism, levels of involvement, you get fewer activists, fewer intellectuals. You get kind of politicians sort of at the top who are making the decisions. And if you think of it in that pyramid, you get different things happening at different levels. So when you look at voters on the Republican side and on the Democratic side, you actually see very similar numbers when it comes to attitudes about AI. People are using it. People who are using it think that it is really, really useful. They get the draw. But there's also real worries. There's worries about job loss. There's worries about, like you were saying, the effect on human relationships. There's worries about privacy. People are have concern and are open to regulation. Typically, you get the stereotype that, you know, Democrats are extremely pro regulation, Republicans are extremely anti regulation. What we found in our research at nrg, as well as, and you can this in other polling data as well, is that people from both sides are interested in having some kind of Rules applied there. You get to the next level up and you get kind of a really interesting landscape of debate on both sides. I think on both sides you have people who, you know, there's these little nicknames for different tribes within sort of the intellectual debate, right? You have doomers on both sides, people who are worried about, you know, the adverse effect that AI could have on people, AI, you know, being outside of human control, AI causing job loss, whatever. You have optimists on both sides, people who think that basically this is going to be a huge unlock in human productivity and is going to, you know, send us to the stars and do all these things that we could never do before. But what you have that's really, really different on both sides is, you know, who is at the top. So on the Republican side, you have Donald Trump, who is basically able, through his, you know, authority within the party, through his role within the party, able to set direction. And what he's done so far is kind of apply his general governing philosophy to a lot of what's happening in AI. So if you hear Donald Trump and JD Vance talk, they are pretty anti regulation. They want AI companies to be able to build, to be able to make their models, to be able to advance with, you know, little, I wouldn't say no restrictions, but with less restriction. They are really focused on China. They want to make sure that the US Is competing with China. They see this as not just a, you know, sort of domestic economic issue. They see this as like a, a global competitiveness issue. And you can see kind of the reverse of the Biden administration when it comes to sort of the social issue aspect of this. So the Biden administration was concerned that AI would replicate all kinds of, you know, racial or gender biases if it was used for hiring, if it was used for mortgages, et cetera. You see of the, a maybe not mirror image, but on the other side you see a concern that large language models will come up with sort of woke responses or socially liberal responses. Right. On the Democratic side, you have no person who has the same poll that Trump does. You essentially have a lot of people coming up with individual ideas, but no unifying philosophy, no governing philosophy. And you, you sort of get this asymmetry here where the Republicans are in, you know, kind of build, baby build mode and Democrats are worried on an individual level, but kind of don't have a way to get on the same page.

A (7:44)

Is there any sense that if Democrats were in power today, what their AI policy would be?

B (7:50)

Oh, that is such a good question. I think there's a couple ways to answer it. You've seen various Democrats emphasizing sort of transparency and trying to get AI companies to say what it is that they're doing to, you know, say what's happening inside of their models, just to get a window in what's happening inside various AI processes. You see, you have a transparency focus. I think there'd be some of that from the progressive side. You see an interest in redistribution. Right. The, the logic on the progressive side runs something like if AI is going to increase efficiency and you know, maybe decrease the number of jobs available, then those AI companies, we should tax or do something with those AI companies so that you can redistribute the gains of those efficiencies across people. We're still in pretty early days of AI. You know, we're what, three years out from some of the early chatbot releases. So it's, and I want to add.

A (8:54)

Here because I was recently looking at this polling, according to Pew Research, this year, only a third of American adults have used ChatGPT. So for the majority of Americans, two thirds of Americans, this is either they're not familiar with it or it's more of a concept than a daily reality.

B (9:10)

Right. Well, this is such an interesting thing, right, because you get, you know, statistics on usage and you have a certain chunk of people who use this. You have a certain chunk of people who are kind of like, you know, on the leading edge, who are super users, who use this all the time. You also get a chunk of people who maybe don't use it, but see it in their day to day interactions. They see it entering their daily life or they suspect that it's there. You know, I've, I've, I've heard anecdotal, I've actually, I've, I've seen this myself. People who a year or two years ago were very sparse with their texts or their emails and suddenly they're sending all this like flowery text around to everyone they know. Right. Even if you don't use AI yourself, you're seeing it pop up in your life, in your world, and you're seeing it pop up in the news as well. So I think it's, it's something where, whether you use it or not, it feels universal. And this is also an economic story, right? You see all these headlines about how, you know, a lot of the, the growth you see in headline GDP numbers is coming out of AI investment. If you look at what's happening in the economy in general, a lot of it is AI. So I think people feel, whether they use it or not, that it's here, if that makes any sense.

A (10:27)

Yeah, I want to start from where we are today before we cast into the future and talk about what crises or not are possible with the use of AI. But like I mentioned, you see on a somewhat regular basis that large companies are laying off, you know, thousands of white collar workers and there's every so often a sort of news cycle that gets caught up in whether or not we're in an AI bubble, whether or not there's going to be a pullback. So I'm curious what your perspective is on where we are today, because I've heard some skeptics say that the economy is in something of a pullback and when executives today want to fire people, they frame it as we're a really advanced, ingenious company that has figured out how to deploy AI in such advanced ways that we can lay all these people off. And so it's sort of creating this like, yeah, this might be bad news in another sort of world, in another news environment, but we're just going to say that we're laying all of these people off or insinuate that we're laying all of these people off because of AI, when really what they want to do is reduce the workforce and AI may not even be playing a big role in it. At the same time, you said that there's this double edged sword of, on one hand, a lot of our economic growth today is because of AI investment. And did we not have that investment, the economy could be doing significantly worse. On the other hand, if we are in the middle of an AI bubble, the burst of that bubble could be particularly painful. So I want to try to orient those two big ticket economic items before we cast forward too much.

B (12:00)

Yeah, well, what you said is exactly the sort of interesting thing and the trouble with trying to figure out causation is because, you know, it is possible that when some companies cut back on jobs that it's just a pullback from over hiring, from the pandemic. That's another one that I've heard a lot of sort of AI skeptics talk about, that this has nothing to do with efficiency. This is, you know, just normal business cycle. What you would've expected to have happened given the massive hiring that happened in, you know, 2020, 2021, 2022. And then you hear exactly the same rhetoric that you're talking about from, you know, maybe sometimes boosters, sometimes critics that say that, you know, AI forward companies or companies that are adopting AI are cutting jobs because they're cutting the lower rungs out from the career ladder and they're replacing people with AI. In my mind, when I'm thinking about the politics of it, I mean, that's an interesting question on the facts, but in political terms, it only matters so much. Right? Suppose you're going into an election cycle. Suppose it's 2026 or 2028 and you're a Democrat and you're looking at the economic numbers and what you're saying is, oh, I don't like these hiring numbers. I don't like the scarcity of jobs. I'm going to say that it's, you know, executives going out and, you know, taking jobs away via AI. I think there's going to be a lot of public willingness to buy that. And I don't think that there's necessarily a referee that can say one side is right, one side is wrong. Right. I think, I think one other way to put this is that the idea of AI job loss is in the discourse. Exactly. To what extent this is normal economics versus AI job loss as a political question, that only matters so much because people are going to see AI job loss as a political issue in the coming elections.

A (13:53)

I mean, do we have some sort of benchmark to look at now for how much displacement AI has already caused?

B (14:01)

I think this is kind of one of the million dollar questions. Exactly. Because of what you were talking about earlier, which is that people have different thoughts about the original initial cause of any job loss. And so some people are going to attribute some job loss to AI, some people are going to attribute it to post Covid restructuring. You hear smart arguments from both sides. So I don't have a single headline metric that I use. What I will say is that people are reporting seeing it in their lives. So this is from a June YouGov survey. They asked, have you or anyone you know lost their job, had their salary cut because of AI taking over some or all part of it? The top line was 1% yes me, 3% yes a friend, 3% yes a family member, 2% yes a colleague, 3% yes an acquaintance. And then a total of. And I think people were allowed to select more than one between those options. A total of 92% were somewhere between no and not sure. So you have a situation where my back of the envelope in my head math is saying that, you know, something like 8, 10% of people are already personally experiencing this or report personally experiencing this. And I think for the politics of it, taking people at their word and seeing what it is they're experiencing is going to be more important than saying, oh, yeah, yeah, yeah, you're seeing this and you're worried about this, but did you know that there's actually some, some other economic cause? I don't think that that that type of argument usually works in politics. Usually the argument that works in politics is, okay, people have lost their jobs, or people are worried about losing their jobs, or people are worried about X, Y, Z, what happens next.

A (15:49)

Yeah, I think that that makes a lot of sense before we get into what potential crises there could be. You know, during the Biden administration, it seemed as though there was some sort of alliance between folks like Amy Klobuchar or Pete Buttigieg, who themselves wrote books that were anti monopolistic, you know, trust and antitrust, and then folks like Josh Hawley who talk plenty about monopolies, and then folks like Tucker Carlson who used their platform to talk about AI and transportation and truckers losing their jobs and things like that. So there was a bipartisan sense that these advanced technologies could cause disruption in American life, unfair business practices that could harm the average American. Do those strains still exist in both parties or have Republicans decided to love AI and sort of cast their lot with the big tech companies?

B (16:43)

Yeah, I think those strains exist on kind of an intellectual level in both parties. You do have people trying to figure this out. But I do think Trump serves as something of an agenda setter here. Right. I think that you, you, I saw reporting about this in Axios about how, you know, sort of the, the populist right left alliance seem be fraying and they had a number of different examples, and I've seen examples myself where, you know, Trump kind of, this is the advantage if you're a party that has a president in office who's sort of a strong decision maker, is that they get to pick the direction that everyone is going. So I think that that generally Trump has been able to sort of set a sort of anti regulation, pro competition with China agenda on the right. I also think that eventually Democrats are going to win another presidential election. Eventually Trump's not going to be around. Right. I think that there's potential. I don't know if that's going to be in, you know, 2029, if there's a Democratic president at 2033. Whenever I think there's potential, when Trump is not such an organizing principle for there to be a little bit more sort of bipartisan work, you can imagine that happening. But yeah, right. Right now I would Say that, that that's the direction it's rowing. And, you know, with Trump in charge of the federal government, states are allowed to make their own AI regulations. That's something that happens. But with Trump in charge of the federal government just kind of letting AI do its thing, I think that's gonna, I think that's gonna continue in terms of being the governing right ideology.

A (18:22)

Okay, let's cast forward. I talked about all of the different areas where AI is touching American life today. From intimacy to the stock market to your car to your home appliances. Where are the potential for crises in which AI will likely be seen as the culprit?

B (18:47)

Yeah, so I've been thinking about this a little bit more, and everyone zeroes in on the potential for some kind of unemployment job crisis where AI vaporizes a bunch of jobs all at once. And I, I get that because I think that's, you know, that's something everyone worries about is how to, how to make a livelihood. I do think that a lot of these concerns, including the job concerns, ladder up into kind of a bigger emotional concern about having control over your own life and having kind of a, a human at the end of the decision chain that can be accountable in some way. Right. So think about all the different concerns. People are concerned about AI going rogue or AI making some, some decision that affects like, you know, your mortgage or a job application or people are worried about AI, you know, having your personal information and doing something with it that you don't want a privacy concern. People are worried about AI weakening personal relationships or, you know, people are worried about AI serving content to minors that's damaging to them. In a lot of these different.

A (19:56)

Driving through an intersection and hit person. Or a cat.

B (19:59)

Or a cat. Right. Like there was that cat in San Francisco that was the beloved neighborhood cat that was hit by a Waymo. And there was massive public outcry. Even though you've seen a lot of data on self driving cars, doing great in terms of safety, you see all these different things that people are worried about. And one common emotional thread is that it's a lack of control. It's that something has happened in your life maybe that you don't like due to a technology that you feel like is a little bit out of reach to you, maybe is opaque to you. I think if we're talking about the public concerns about AI, that's one of like the big drivers at the heart of it. And I don't think anyone has fully figured out how to address it. Right. Like the, the strength of The Republican plan is, you know, suppose that the optimists are correct and AI just like shoots human life to the moon. Maybe metaphorically, maybe not metaphorically, but like it's just this unalloyed good for people. Then, you know, the strength of the Republican plan is that if that is the case, then they've sped it along, they've done something that's good for society, they've done something that they'll get broad credit for, they'll get credit as being farsighted, et cetera, et cetera. What's missing is that second side of the equation, which this often happens in our two party politics, is one side figures out one thing to or one aspect of an issue and then the other side figures out another aspect of an issue. You know, what's missing is that I think Democrats haven't, haven't landed that plane on how to address the worries, address the feelings of lack of control that some people have around this technology.

A (21:43)

Yeah, and I want to get into that a little bit more. But can we be specific about what kind of crises could make Americans feel like they've lost control of the machines?

B (21:57)

That's a really great question. This is one where also we're in such uncharted territory right now. If you think about a lot of the other technologies that have existed in human history, they haven't really resembled AI. A lot of them are not necessarily, you know, knowledge creation or knowledge aggregation or decision making technologies per se. Right. People compare AI to the steam engine, but the steam. There were still people operating steam engines. Right. People create. People compare AI to the Internet in terms of its impact. And the Internet is still a bunch of people sitting behind computers. Right. So I think job loss is one to keep an eye on. And I'm agnostic about the whole bubble question personally. I've tried to know what I don't know, and I know that I don't know a lot about economics. And there's so much speculation involved in the bubble question. I think if a bubble did exist and it did pop, that would be something that would cause some kind of crisis. I think that if it's one off incidents like what we're talking about, like a cat getting hit by a car, I feel like that doesn't necessarily hit the bar for creating some sort of crisis. I guess you could imagine something geopolitical. I don't have enough insight into like what goes on in the national security apparatus to know if there are actual like core decisions being handed off to AI. I would assume not. So I Think probably the economic realm, probably job loss. If I'm thinking out loud, you could imagine another scenario where it's bit by bit, it's piece by piece, there's some job loss here, and people are seeing more AI generated content that they, that they dislike and they don't like that. And it kind of. Where you kind of see it chipped away. You can imagine that. I think, yeah, but we're, we're in uncharted territory where also I, I think, and this is something that is, is a little bit speculative on my end, it's something I haven't done, like, deep research on and that I want to do more research on is when people are thinking about something bad and there's a person on the back end of that bad thing versus a computer on the back end of that bad thing. Like what happens? Right. Like if a bunch of people lose their job in a financial crisis that's precipitated by a whole bunch of banks doing all kinds of bad trades and using all kinds of vehicles that they shouldn't have used. Right. Do people process that differently than if an AI bubble bursts or if their CEO, in order to stay competitive with a bunch of other firms that are doing the exact same thing, vaporizes a bunch of jobs and the new jobs that are created are not a fit for them? Right. There's all those kinds of crises that you could imagine, and it's a little bit in a day by day, wait and see mode to see if any of them materialize exactly how it looks. I know that's probably a hugely unsatisfying answer, but that's a window into where my personal brain space is, if that makes sense.

A (25:10)

Well, I think the reality is from a political perspective that there will be people at the end of the chain who are viewed as responsible or who are responsible for whatever the crisis may be, whether it's because they run the companies or because they were charged with regulating the companies or whatnot. And one of the things that you said to me in your stream of consciousness about politics and AI was we are in this moment where the Democratic Party doesn't quite know what it is or what it wants to be when it has the opportunity to present itself as a unified force in the next presidential election. And that's gonna require some innovation or will likely involve some innovation. And you pointed to past moments where parties were in this, you know, they'd been in the wilderness or felt that they, you know, Democrats were just in control. But, like, I don't think that the Biden administration felt particularly satisfying to a lot of Democrats. And so Democrats feel like they've had a hard decade of it and they want something new. They want to turn a page. And, you know, maybe you can talk a little bit more about how parties have handled this moment in the past before we get into how they might use AI to reinvent themselves.

B (26:27)

Yeah. So this is so interesting. I'm gonna. One book that I think about a lot, because this is who I am, is a book called Party ideologies in America, 1828-1996. And it's this book that essentially makes the argument that in different eras, you have kind of governing big picture ideologies for parties. And after a while, when a party accomplishes enough of its goals or when conditions change enough, when time just moves on enough, a paradigm becomes sort of expired. So a good example of this right now is the Republicans sort of pre Trump versus the Republicans post Trump. Right. You could say that from sometime in the, you know, 70s to 80s through, say, 2015, Republicans had a pretty clear paradigm. They had foreign policy hawkishness, they had economic libertarianism, and they had, like, social conservatism that was grounded in sort of very religious issues. Right. That was when same sex marriage was sort of a headline issue for the Republican Party. That was when abortion was a headline issue for the Republican Party. And those things were able to sort of bring together camps. But, you know, eventually, over time, you get enough successes and you get enough failures that that combination no longer works. Right. Like Bush and Reagan cut taxes. The Cold War is successful. The war in Iraq, by many accounts, is not right. And on social conservatism, abortion is essentially, you know, prior to Dobbs, sort of fought to a draw in a lot of ways. And same sex marriage becomes the log of the land. And you have a situation where a paradigm kind of expires. I think Democrats are in that kind of moment. You could argue that somewhere in the 90s to mid 2000s, they figured out a kind of Democratic universalism. That's at least the term that, you know, this guy John Gehring uses in the Party Ideology book, where essentially what you're able to do is you're able to stitch interest groups together and kind of create a sort of coalition with a lot of different people that's premised on, you know, social liberalism, on kind of technocratic expansions of the welfare state and a kind of multicultural affect that just kind of works from Clinton through Obama. But I think what we saw in, you know, Hillary Clinton's loss in Joe Biden's, you know, victory, but then kind of the subsequent unraveling of his presidency and then Kamala Harris's loss is that there's something about that Obama paradigm that is not responsive. Right. Like no paradigm lasts forever. And I think Democrats are kind of trying to figure out what's their big approach that weaves together their disparate instincts that weaves together all these groups that have some things in common with each other, but some real differences between each other. Like what's that next thing? I think that's, that's the fight of the 2028 primary. That's a fight that I think, I think honestly got put on hold for a long time because Democrats were so unified by how much they, they, you know, hate Trump. But you know, Trump's, he's, he's, he's won twice. The Republicans are going to have somebody else in the next election running. They've got maga, a MAGA sort of ideology that does not feel spent, that feels like they have, you know, other goals and ideas that they want to pursue. And Democrats haven't figured out their response yet. So I, 2028 is an opportunity for that. But sometime soon I feel like a pivot is due, if that makes sense.

A (30:12)

Yeah. And I think folks are pretty familiar with how Republicans pivoted in the aftermath of Reagan's three legged stool and it was on immigration issues and American identity and pivoting to economic populism, at least in name, if not necessarily in practice. I think Democrats have started to get a sense of what they want, which is, you know, the little guy versus the billionaire, the sort of corrupt state versus the everyday American. And you don't have to have a big imagination to see how AI could fit into all of this. Particularly if we find ourselves in either an AI crisis or a crisis that is perceived to have been the result of AI, which as you suggested early on for political purposes, it probably doesn't matter all that much which one it is. So how do you see those two things potentially colliding? Democrats in search of a new paradigm and the possibility, well, the possibility of an AI driven crisis. But maybe we don't even have to get that dramatic. Yeah, that there are lots of human insecurities, worries, concerns, vulnerabilities that are provoked by technological change. And whether we have the full on meltdown or not within the next three.

C (31:34)

Years, I think the amount of attention.

A (31:37)

That we pay to this issue only grows.

B (31:41)

Absolutely. Amount of attention paid only grows. And I think Democrats kind of have to be prepared for a couple different scenarios, right?

C (31:49)

All right, that's the end of today's preview. Head over to GDPolitics.com to become a paid subscriber subscriber and listen to the full episode. We chatted for over an hour and if I'm being totally honest, I think this was one of my favorite conversations that I've had since we started GD Politics. We get out of our pretty tired current political moment and talk about what could come next.

A (32:11)

We talk about what parts of AI.

C (32:13)

Disruption could be mined by ambitious politicians and how Americans feel about it all. We also talk about the bots that are now answering polls and even the value of authenticity in politics. Like I said, head over to GDPolitics.com to become a paid subscriber and catch the whole thing. Paid subscribers get about twice the number of episodes, can join in the paid subscriber chat, and most importantly, keep this podcast going. When you become a subscriber, you can connect your account to wherever you listen to podcasts so you'll never miss an episode again. Head over to GDPolitics.com, see you there.