This transcript is generated with the help of AI and is lightly edited for clarity.

SARAH LONGWELL:

Why didn’t you tell them they were wrong? Why didn’t you push back on them? Why didn’t you correct them? And I say, “Look, I sit and listen to 10 people so that I can figure out how to persuade 10,000. And it is not my job to tell them they’re wrong. I want to know why they think the thing that they do. I want to know where they got that information.”

REID:

Hi, I’m Reid Hoffman.

ARIA:

And I’m Aria Finger.

REID:

We want to know what happens, if in the future, everything breaks humanity’s way. What we can possibly get right if we leverage technology like AI and our collective effort effectively.

ARIA:

We’re speaking with technologists, ambitious builders, and deep thinkers across many fields—AI, geopolitics, media, healthcare, education, and more.

REID:

These conversations showcase another kind of guest. Whether it’s Inflection’s Pi or OpenAI’s GPT or other AI tools, each episode we use AI to enhance and advance our discussion.

ARIA:

In each episode, we seek out the brightest version of the future and learn what it’ll take to get there.

REID:

This is Possible.

ARIA:

The past few years in American politics and the lead-up to the 2024 presidential election have been wild. Regardless of your personal politics, I think we can all acknowledge that America on the whole has been navigating deep divisions between parties and within parties. Divisions within businesses, campuses, congregations, family dinner tables—everywhere.

REID:

Possible is all about envisioning the best possible future, often from the lenses of renowned technologists tackling challenges. To help us get there. Today, we’re going to talk to someone who is an expert in a different type of challenge: understanding the political and societal divisions Aria mentioned. So why go here? Because another principle I hold close and deem absolutely critical to business innovation and humanity is democracy. No matter who wins, no matter where we disagree, upholding democracy matters deeply.

ARIA:

I also think it’s important to clarify something. This isn’t an episode that’s meant to proselytize or convert anyone. Listeners, you can probably pretty easily guess who Reid and I are supporting in the upcoming presidential election. But our hope here is to better understand something we’ve been thinking a lot about. That is what exactly is at the root of this division and the threats to democracy that we’re seeing today. We are talking to one of the smartest people we know when it comes to listening to American voters and getting to know them.

REID:

Sarah Longwell is a Republican political strategist and pollster who co-founded Defending Democracy together in 2018 and Republican Accountability in 2020. These organizations combat misinformation in politics and support some Republican members of Congress as well as a number of Democrats who defy the growing trend of anti-democratic sentiment.

ARIA:

Sarah also co-founded The Bulwark, a news organization that advocates for democracy through analysis, reporting, and podcasting. She’s a leading force in the pro-democracy ecosystem and an expert in how people consume information and make decisions. Here’s our conversation with Sarah Longwell. Well, Sarah, so lovely for you to be here. And so one of the many things that you do is you are in the business of running focus groups. And at the top of the interviews you like to ask participants, “How do you think things are going? What’s important to you?” So I’m going to turn the tables: How are things going and what is important to you right now?

SARAH LONGWELL:

Well, look, I mean, the way things are going is considerably better than they were going. And really, when it comes to voters, the thing going into this election that was causing me a lot of anxiety, because I am somebody who believes that Donald Trump represents a unique threat to the country, was that Joe Biden wasn’t going to beat him. And it’s because I’ve been listening to voters for the last couple of years talk about their concerns about Joe Biden’s age. Like, it just was so people were so unhappy with this choice that when you would listen to voters, especially the ones that Joe Biden both needed to win over — so persuadable voters, swing voters, people maybe who voted for Trump in ’16, but then had voted for Biden in 2020, and so they were part of the coalition that put him over the edge — those people were just, they were backsliding toward Trump.

SARAH LONGWELL:

They were frustrated. And when the change happened, I have, I got to say — I was a, I was a proponent of Joe Biden stepping aside after that debate. But even I, as vociferously as I felt that, I was concerned about Kamala Harris, because I had been listening to voters sort of present a negative impression of her for a long time. And I am as surprised as anybody to have seen how the Democratic coalition has coalesced around her and just decided we are going to be pumped. And so the thing that has changed is the entire tenor of the electorate that ranges from sort of base Democratic voters to sort of soft Democratic voters — and even some sort of persuadable swing voters in terms of their level of enthusiasm. And that enthusiasm gap was what had me the most concerned. Because the thing about Donald Trump is that while he is a repellent to a certain kind of swing voter, and he is a turnout mechanism for Democratic voters who dislike him, he always has a base of support that is very enthusiastic about him. And so that — the chasm that was opening up on enthusiasm just had me, had me worried. And so today, now that we are sort of in this space of, it’s just a world of possibilities.

ARIA:

Can I just ask the question that I feel like everyone says to me? They’re like, there are no undecideds. They’re like, how could you be undecided? We know these two candidates, or at least these two parties so well. So you really believe that there are undecided voters and so there is persuasion work to be done, sort of things like that, because I think people are skeptical.

SARAH LONGWELL:

Yeah. And the reason that they’re skeptical is because they themselves are high information voters. Right? Anybody who’s listening to this podcast is interested in stuff. They consume lots of these things. Anybody who listens to one of my political podcasts over at The Bulwark or any, they are the ones — these highly engaged political podcast listeners. They’re the weird ones, right? They’re the minority, right? But those are the kinds of people who don’t believe that anybody could be undecided. But what they — I think don’t think — realize is that people say, use the word low-information to talk about a certain set of voters. I think that those people are just normal <laugh>. They’re just normal people who are busy living their lives. Totally. You know, they got kids to raise, they got fun stuff to do, they got, or they got jobs that are hard things that take up a lot of time.

SARAH LONGWELL:

They got sick parents, they got volunteer things. They are not sitting constantly. Like if they’re on their phones, they’re scrolling through headlines, whatever. And so it’s not that they’re undecided, like, “Boy, I’ve thought really hard about this and just can’t make up my mind.” It’s that every four years, because a lot of these people don’t vote in midterms. They vote in a presidential election. They believe it’s important to vote, but they look up, you know, a month out and go, “All right, I’ll look at these two, you know, these two evils, and which one is the lesser of two evils, and which poison am I drinking?” Because they tend to have a pretty cynical view of politics. And so that is what an undecided voter looks like. It looks like somebody who is a late-breaking attention payer to the political landscape.

REID:

You know, one of the things that — you know, my own getting into the political arena, which is mostly because I, you know, share your point of view that, you know, Donald Trump is a unique corrosion. I find such entertainment — “Oh, oh, he is the business president.” It’s like, yeah, he is the business president who bankrupts all of his businesses. Like, it’s like, you want business people in the various instruments of government, but ones who are good ones. Or at least even B- ones is okay. But say a little bit also about like, some of the dynamic that tends to be the, do you motivate people enough to go vote like in the polling? And do you, do you persuade the other side that go, “Well, none of my family’s voting Democrat we’re kind of like, we don’t vote Democrat. Not our thing, but we’re going to stay home because we think Trump is too, you know, too, just too toxic and too fugly about himself,” [laugh] and so and so. Yes. It’s a technical term. Yeah. So how does the polling mechanism actually inform us? Because, you know, one of the things that I’m amused by is like, people go, “Oh, and look, the polls moved 1% today.” And it’s like, well, I tend to not pay attention to that stuff.

SARAH LONGWELL:

Yeah. I mean so obviously I do focus groups in part — and I exclusively do qualitative, because I find quantitative to oftentimes not give me the information that I really need. And one of the things just to sort of back up to some of the, one of your earlier, the way you were framing it, I would say among people like me who are operatives in the space, thinking about actually how do you win an election? There tends to be almost like a debate between turnout or persuasion. And there should never be a debate about that because you absolutely need both. 2020 is a perfect example, both sides — because Donald Trump is so polarizing — absolutely turning out every last voter that they could get their hands on. And then the race, the election being decided by sort of a critical margin of the soft GOP right-leaning independents in places like Pennsylvania, Michigan, Wisconsin, Arizona and Georgia.

SARAH LONGWELL:

And like that is, that’s it. That’s the ball game. It is turnout plus persuasion. You need them both. But the enthusiasm makes an enormous amount of difference to Americans. So in every election there are big things at play that are part of the human psyche. And one of them is, so when people think about how to motivate people, there’s five dominant human emotions: fear, love, anger, greed, and sympathy. And the ones that motivate people the most are fear and anger, right? And so I actually think for too long, in a way that is — to use your word — corrosive, the fear and anger that Donald Trump generates has been the turnout mechanism. And I think that that is causing — there, there’s some fatigue in that, right? And I think that what Kamala Harris coming in has done is it has allowed people to have someone to be for, which is why you’ve seen surges in registration, surges in enthusiasm, surges in donations, surges in people who are volunteering.

SARAH LONGWELL:

Those metrics mean a lot more than polls. Because when you have enthusiasm in a human experience kind of way, instead of people saying to each other, “I don’t know, man, I’m going to, obviously I’m going to vote for Biden, you know, but God is he old and I hope he can, doesn’t fall over the next time he gives a speech and whatever,” like that person who’s saying that, okay, they’re going to go vote because they hate Donald Trump, but is the person they’re talking to who’s much less interested in politics or maybe like cares less, does that motivate them to go vote? Or when the person’s like, “Oh my gosh, I am so pumped to vote for Kamala Harris. I love how she’s talking about the middle class. I love how she’s a prosecutor,” or like, whatever it is. Or I’m just, “I’m high on vibes, I’m huffing vibes.”

SARAH LONGWELL:

Like, does that translate into the person that they’re talking to who’s less interested in politics being like, “Yeah, I’ll go vote for her. Yeah, I’ll do that. I’ll, I’ll take some time out of my day,” and I think that that is something that just can’t be captured quantitatively, which is why I love talking to voters. because I want to hear the tone of their voice. I want to hear the words they’re using. I want to know if they’re being dragged kicking and screaming to something. I want to know if they’re voting against or voting for if it’s fear and anger motivated. And it’s always a little bit of that. But like, is there also something positive? Is there some love in there? You know, is there something, something to — and because once you’ve got something to be for and something that you’re really against, man, is that a potent combination.

ARIA:

I consider myself a data person. I, I, yes, I fervently check the polls. I’m like, give me the numbers on that. I was just talking to someone just this week about a focus group and they were like, “Oh yeah, but that’s just a focus group. Like, who cares what 10 people think?” So can you tell us like, how do focus groups actually work? How do we even find these people that either say they’re undecided or, you know, were Trump/Biden voters. Like give us the run through about how you do that, because I think some people are skeptical.

SARAH LONGWELL:

I love answering this question, because I think it is a mystery to some people, and it’s actually very simple. There are places that are firms, that they have over the years collected massive lists of people and they know tons and tons about these people. Because most of the time when people are running focus groups, they’re actually not talking about politics. A lot of times focus groups are like, “Is the cat food that you buy, do you like it better in a blue bag or a yellow bag? Which one grabs you from the shelf?” Or “Here’s the new menu from Chili’s. Like, does the chili double cheese fry bacon thing sound delicious to you? Or disgusting?” You know, like people use these lists to talk to people about all kinds of things when they’re seeking deeper information and they’re trying to understand not just the what, but the why. Why does somebody feel this way?

SARAH LONGWELL:

We go to those firms and I’m like, “Okay, firm, could you please find me people who voted for Donald Trump in 2016, Joe Biden in 2020 and are undecided right now? And I would also like them to be college educated and I would also like them to be from Pennsylvania.” Like, you know, when you give them sort of a deeper and deeper cut. And I think that a lot of the ways that people use focus groups in politics, it is oftentimes campaigns and they will do a series of focus groups about the candidate, and they’ll do like six and they’ll do a bunch of different demographics. They’ll do a group of women, they’ll do one group of men. They’ll do college educated, non-college. And they’ll slice and dice a couple ways and then they’ll, they’ll learn what they learn and they’ll move on.

SARAH LONGWELL:

The reason I started doing focus groups in the first place is that back in 2017-ish, late 2017, we were like into the Trump administration. And I was a sad Republican who was like, “What are we going to do about this guy?” And my biggest idea was, well, we got primary him. Okay, guys. And I’m like running around to governors and congresspeople that I like, Larry Hogan. I’m like, “Larry Hogan, you must primary Donald Trump.” And you know, anybody that I talked to was sort of like, “Yeah, maybe. Well, but how about you show me that it can be done? You show me there’s an appetite from voters.” And the thing, one of the key mistakes that I was making at the time is that I felt like I was a representative of the Republican party. I thought people were, I — the people make this mistake all the time in punditry, they substitute their own judgment for the judgment of everybody else kind of in their tribe or whatever.

REID:

Oh, I never do that [laugh]. That’s—.

SARAH LONGWELL:

It’s an easy thing to do, right? You’re like, well, “I speak for all tech, you know, titans who you know…” There’s, and so I felt that way. I thought surely okay, that Donald Trump was an accident of history. That it was the product of a fractured field. He is over there all alone in his burn-it-all-down lane. And everybody else was splitting the normie vote. And so this, most people felt like I did and they wanted this guy gone. So I started doing some focus groups and some polling to figure out if he could be beaten. And immediately I realized the answer was no. I realized that there was an affirmative interest in people who were there for the burn-it-all-down. There for the chaos. There for — that they saw Donald Trump as a businessman that was going to help the economy.

SARAH LONGWELL:

We all learned a lot of these lessons in 2016, 2017 around, but we’re like you can’t say that, “Grab ’em by the lady parts” and win an election. But of course you could. You can’t say Mexicans are rapists and thieves and win an election or talk about a Mexican judge. Oh no, you can. You can’t talk about a Muslim ban and win. And it’s like, actually your, his poll numbers went up. We were learning in real time things that we didn’t think were true lessons about the American electorate. In doing the focus groups, I realized like there was a real scales from my eyes moment of being like, “Okay, I’ve spent too much time in think tanks talking about education reform and debt and deficit policy.” And you know, I got like addicted to knowing, because talking to voters is cheat code for understanding American politics.

SARAH LONGWELL:

Because that is who drives in a democracy, the political environment. What I found was that if you did focus groups every week, which is sort of the cadence that I started to do, you would hear themes. You would hear voters say that — you’d hear a voter in Pennsylvania and a voter in Arizona. And you would be like, “Okay, they’re talking about this the same way.” And you could start to decipher what was in different demographics, what were sort of key indicators of the way somebody might think. You could understand which media — I mean, one of the questions I ask all the time is what media people are consuming. When the pandemic hit, I was like, “Oh God, I have to keep talking to voters. How am I going to keep talking to voters?” And so I went to all of these focus group facilities and I was like, can we do this by Zoom?

SARAH LONGWELL:

I’m doing everything else by Zoom now. And it was like a miracle. You’re suddenly teleported into all of these people’s living rooms. And sometimes they’re smoking, sometimes they’re vaping, sometimes their babies are crying, their dogs are barking. But like you still get 85% of the nutritious feedback that you need from them. Like you miss out on a little bit of the dynamic in the room and being sure about some of the ways that people are giving you facial expressions. Like you just miss a little bit of color, but you get still most of what you need by doing it by Zoom. Which I was like, “Oh, oh, we can do a lot of these now.” A lot of times, I’m just like, “I need more focus groups. You have to feed my focus group addiction.” My greatest most fervent hope is to find a way to make qualitative more scalable.

SARAH LONGWELL:

Because I think it is much denser and more interesting and more actionable to have real people’s opinions as opposed to just a number. This many people think this, and this many people think this. Because every time I see a question like, or an answer like that, I’m like, “Why was the question written? How was it worded?” Because I know from listening to people over and over, that two words can change the entire meaning about the way that somebody answers a question. And so I want to know how somebody was asked a question and how they answer it in order to understand what I think is the true meaning of what they’re trying to get at.

REID:

You actually anticipated where I was going to go with my next question, which is the role of technology in evolving how we come to the understanding of what do — for example, in politics or in business, you know, blue or yellow — do we understand kind of like, as it were, the will of the people or the interest of the people and the nuanced shape of it. And Zoom, this transformative force where the technology and the pandemic came together, is obviously one. What are the other areas where kind of technology giving us kind of a better lens into try to having a, you know, kind of like understanding our human interests, our human needs, our collaborative will, our ability to kind of navigate conversations with the body politic? What are some of the things that you do? What are some of the things you’re hopeful for? How do you look at the evolving technology of this?

SARAH LONGWELL:

So yeah, I’ll give you a couple examples. So one of the things, when you used to do the focus groups in a room, right? And you had people sitting there, there’d be a camera that would record it and it would be like kind of a blurry, like, you’d get like this blurry image of everybody’s head at the top and the audio’s kind of garbage, because people don’t have like good mics in front of them, whatever. When we went to Zoom, suddenly we would have transcripts — like, you can use the AI stuff to do really good, whether it’s Descript or Otter, something like that, you can get a really good transcript. Okay, well now you do hundreds of focus groups and you got transcripts for all of them. And so my, when I think about my hope for what you can do it is, can you take all of those transcripts and put them into the machine learning that allows you to ask questions around trends that have occurred, right?

SARAH LONGWELL:

So you say when we ask about immigration, “What are the top three answers that we get from two-time Trump voters?” And have it come back with, right, a basic summary of what you’ve heard from a lot of these people. That is like the most basic thing that I want is to be able to take the transcriptions and be able to almost make it — not just searchable — but able to get answers to what people have said over time and maybe the way that answers have changed. What kind of words have they used? That’s the first thing that I would say I’m, I’m interested in doing that. We’ve, we’ve sort of been experimenting with. The problem is when you do a two hour focus group, a lot of words are generated — a lot. And so to have, you know, many hundreds of those and then make it searchable, that’s a, you need a lot of — the machine needs to be able to take in a lot of information.

SARAH LONGWELL:

So that’s the first thing. Then the second thing you could do that I’m very interested in is like, can you make it predictive? Right? So once you’ve fed all of this information and it’s got hundreds and hundreds of different voters talking about things, can you ask it? So for example, actually, I’ll give you one right now. So we’ve asked about RFK in all kinds of focus groups and we were like, “How are two-time Trump voters going to feel about RFK dropping out? And like, will they vote for him?” And it will find for us all of the people who’ve talked to them — because lots of people in a focus group that we may not be able to go back and find ourselves are like, I’ll remember. You know, there’s one guy who said if RFK dropped out, that he would definitely vote for Trump if that, who said that?

SARAH LONGWELL:

How many people said that? Whatever. Like, you could ask it that and it could tell you things that for me would be actionable with reporters, with my analysis. Okay? That’s another way that I think it could be really interesting. The third way, and the way that I think I’m the most excited about for the future is, can you ask one question, the same question to thousands of people and get the answer back in words, right? So this is like polling, but in polling, right, you can do yes or no because you can’t, there’s no way to sort of take the answer in — but can you ask it, the same question, to thousands of people and then get the summaries back with direct quotes, but also like word clouds and what were the things that people used the most and make their answers searchable. Like, there is a way to get qualitative so much closer to quantitative from a scale standpoint, but then so much richer in the density of the answers that you’re getting with the human inflections. And to understand, you know sort of the more, the richness that’s in there. And that is the part that I get the most excited about, because the only thing that I want is just more access to more people and more qualitative data and better ways to analyze it.

ARIA:

You know, Reid and I talk a lot about both being truth tellers, but also engaging with people who disagree with us. And I feel like I have a bad quality of being like, “Yeah, me and that person agreed. Didn’t you just hear?” And someone saying like, “No, no, no. Did you hear the subtext? Or did you hear when they said x, y, z?” Like, do you have tips for like, how do you actually listen to someone? How do you tease it out — what they’re actually saying, just so that we can all be better listeners and perhaps better persuaders?

SARAH LONGWELL:

Oh my God, I love this question so much. Because here’s the thing actually, the biggest downside, if you’d ask me sort of a different question of like, “Well, what, what’s the most — the biggest downside of qualitative, right?” And I would say, it’s some version of you, you, you’ve — sometimes you can hear what you want to hear from people. And you, you do have this sense of yes, like you filter people through your own frame. 


ARIA:

Totally. That’s what I do. 

 

SARAH LONGWELL:

No, also, and also you have to be aware of how the people in the groups impact each other. Right? So like, you have to be able, and this is where the human element — I’m not sure you can replace this part of it — but it’s actually why I like the idea of asking people the same question outside of the room together.

SARAH LONGWELL:

Because the biggest downside to the qualitative in the room is that people are having an effect on each other. There is always a group dynamic where people want to get along. And I’ll tell you one of the ways that we tried to do an end run around this early on, one of the things we used to do — people will ask like, “Why don’t you have a bunch of, you know, some Trump voters and some Biden voters or some Harris voters or Trump voters in the room together?” And we realized pretty quickly actually that was not productive because people didn’t trust each other. And so they would either argue or they would, they would change how they answered things in order to make it — you know, they do a lot of like, “I don’t want to offend anybody,” or I don’t want — or throat clearing.

SARAH LONGWELL:

And so it’s clear they were adjusting themselves for the people in the room. When we do a group, we tell them, the first thing we tell them is, “You all have a lot in common. You all voted for Trump twice. You all live in swing states, you all are, went to college.” Like, and you know, we tell them the things that make them say, “I’m safe, I’m safe with my tribe here.” And they, and because they are more honest with you when they feel like they are around other people who are like — and my best moderators are the ones who are on everybody’s side. People get so mad at me when, because I have a podcast called The Focus Group and we play the clips of the people talking. And people are so enraged by what people are saying and they’re like, “Why didn’t you tell them they were wrong?

SARAH LONGWELL:

Why didn’t you push back on them? Why didn’t you correct them?” And I say, “Look, I sit and listen to 10 people so that I can figure out how to persuade 10,000. And it is not my job to tell them they’re wrong. I want to know why they think the thing that they do. I want to know where they got that information.” And so the moderators, they nod, they don’t confront. Now sometimes they push back, they’ll be like, “Well, what if I told you x?” Or you know, like there’s, there’s prodding, there’s not no engagement, but you are trying to get people to answer as honestly as possible. Because the other thing people worry about is they, they worry about being wrong, right? They suddenly feel like they’re being asked a test. And so we want to make sure that they feel safe to say, “I don’t know anything about that.”

SARAH LONGWELL:

If that’s the case. Because one thing I like to report on all the time is: Did people see this thing? I mean the biggest lesson I’ll take away from talking to people is, and it goes back to something we’ve already talked about previously, is how much people are not paying attention the way those of us who are doing this work are paying attention. It’s like there was a time, the time that Trump went to Helsinki and got on the stage with Vladimir Putin and sided against America’s intelligence community. I will never forget walking into a focus group in Ohio and being like, “What did people think about what Trump said in Helsinki?” And everyone was like, “What are you talking about? I have no idea what you’re talking about.” It was like there was a massive conflagration in D.C. and in the political media, none of these people had heard about it.

SARAH LONGWELL:

They didn’t care. They’re not listening. And when I, when you, when you start to understand that, right? And, and you give them permission to not feel like they have to give their opinion if they don’t know something, that’s valuable information too. because You want to know what’s breaking through, what’s not breaking through. You do have to be aware of all of the factors that are impacting the way people are talking and be able to discern — and this is just deeply human — because you have to discern when somebody seems like they really mean it. I watch in the groups when we ask, “Okay, who’s — if election’s today — are you voting for Harris? You know, hands up. Voting for Trump, hands up.” The person who goes last and like waits for everybody else. You’re like, “I don’t, I’m going to give that person a half. I’m not sure they meant that.” There does require some judgment around human nature and behavior and you try to correct for as much as you can, like making it safe. And then some of it you also just have to be able to like, I don’t think that’s a strongly held opinion. And that’s why you never want to take some outlier kind of opinion and be like, “Well, one guy said this,” and then substitute that for a political trend or a way a lot of people think.

REID:

How do you think this will compare when we start having AI driving focus interactions? Because we, with AI, we’ve essentially passed the Turing test, in terms of language. You know, getting to video is still a challenge, you know, but audio is real time now. And so you could do audio conversations or you could do one-on-one mass. And to some degree, you know, I tend to think you’re going to get an interesting lens because that pure individual versus people feeding off each other and a little bit of like, you know, kind of like talking to the — and you go longer and, you know, do more time trust building, you know, have the thing always be going. Yeah. Yeah, that makes sense to me. You know, as if the world is flat. Yeah. Yeah. That makes sense to me. Because this is an inevitable part of the future for how this is going to play. And what do you think, what do you think is going to happen? What do you think we should do? What do you think we should steer away from? What, what will be an intelligent use of this?

SARAH LONGWELL:

Well, I mean, like I was talking about before, I have my own ways of thinking about how it could be applied to the work that I do and other people do, who are trying to sort of information gather around people. And I think you sort of square the circle about a couple of the things that I talked about, which is, “When do you need a human being to be able to discern something about other human beings?” And when are you able to take what the AI can do and let it aggregate things for you that you otherwise can’t aggregate because it’s too onerous. But I’d be interested actually in — you’re listening to me talk, and you are an expert in this, I don’t can’t think of all the ways it could be used. I mostly think about, as a practitioner of this, like what do I want more of?

SARAH LONGWELL:

The thing that I don’t know about AI that I think is a challenge with human listening is the tone. Like, was this said in a skeptical way? What’s a joke and what’s not a joke? When was something said enthusiastically and when was it said with — if there’s not a verbal qualifier, like “I’m not sure,” or “I’m ambivalent about this,” — how do you read ambivalence in somebody’s voice? Because I’ll tell you, when I watch a focus group, I can read ambivalence in people’s voices or when I don’t think they’re sure. And so I think that that’s the part I’m most interested in — is the key to the qualitative — is the human element.

REID:

Well, so the thing that’ll be interesting is I, in the near term — call x years one, two, three, you know, very hard to make predictions about AI further than two years out in a lot of like, not because, “Oh my god, it’s terrifying or amazing,” it’s different shape and lot to evolve. It probably won’t be as good as the best human experts, but there’s kind of three things that can obviate that. One is they’ll probably be pretty good at saying, “Hey, this is one I couldn’t read that. Well, I’m going to punt that to a human.” So you get what — in AI we describe it centaurs: human plus machine performance — which I’m kind of broadly optimistic about. Like that’s going to be a decades long and possibly longer kind of part of this trend, which is humans are going to be humans using AI, you know, doing a whole wide variety of these shops.

REID:

That’s one. Two is, you’ve also got the overwhelming data, right? So like it’s very expensive to get the expert human to do it in each of these cases. Whereas like you’re going longevity, you know, say for example what you did is you say, “Look, well, you know, the way you compensate focus group people, you pay them for the time. So look, we’ll pay you for your time, but what we want is 15 minutes every week, right?” You know, at Friday at the end of the week or Monday at the beginning of the week or whatever the, the, the thing is, and 15 minutes of that conversation. And then you get, you know, kind of paid at the end by doing the whole thing. And you have the longevity for each person with this. And you can do that in a way where you generate, so the, the raw amount of data that you’re going to be able to generate to start looking at it.

REID:

Now, that’s part of the question. You have to kind of look at it and look, look, is there a systematic bias in the generation of the data that becomes, you know, something that’s, that’s careful too, but the question about being, for example, sympathetic and human in, you know, and prompting, that stuff we can already do. And could do well. Like for example, you know, part of the optimism use case that I see about AI is like having an educational tutor that is infinitely patient on every subject for every age, right? And is trying to figure out how to help you learn this thing. And the AI is going to be a lot better than most human beings at that because, as opposed to, “Oh, am I explaining algebra to you again?” It’s like, “Oh, let’s try that again.” Right? You did a really good job last time.

REID:

But there’s a few things that we want to, we want to work on in order to get to, in order to make it happen. And, and, you know, I think as I, and will change in these economic-landscape ways, is that like your both volume of data and your ability to interpret and deal with volume of data is like here now. You know, I don’t think anyone’s using AI as the driver for focus groups yet that I know of. But, but I think it’s inevitability and I think there will be strengths and weaknesses about how human beings, like, “Can you kind of read the minds?” as it were, through the language, making them feel safe. Say, “Oh, I can speak my truth about, you know, why it is, you know, I think, you know, Trump’s a deranged narcissist.” And I can do that here versus, “Oh, I don’t want to create any, any, you know, kind of ill will or, or kind of arguments in it.” One of the things that’ll be interesting when it gets to having AI is in some sense AI will make that more, more easy because it’s like, “Well, it’s kind of like I’m talking to my computer or something else.” It’s like, I don’t have, I won’t have the same protections but on the other hand, I’ll have other biases. And I’m trying to figure — like, I don’t know yet.

SARAH LONGWELL:

I can see how separating people out from a community of people would allow them to be more honest in certain circumstances about things. But let me give you an example of where it works in the inverse. So one of the things that we ask about a lot is abortion. And one, and one of the things that I always find super interesting is when I ask people, “How do you think things are going in the country? What issues matter to you most?” Which is the way we kind of kick off every group, people don’t say abortion that often. And that’s always really notable to me because when we at some point introduce it — so we say, “You know, no, nobody brought up abortion. How do people feel about abortion?” The level of emotion that you get from people then, as they engage on that, is extremely high.

SARAH LONGWELL:

They’ll be, oh, and you know, suddenly they’re leaning in and they’re having a conversation. And especially if you have a group of women, the same thing always happens. People start telling stories about themselves or their friends where they had a catastrophic pregnancy, they themselves had an abortion, somebody they knew had an abortion. And a tribe gets formed in the moment, where the women start to support each other, start to talk to each other with great empathy. And I think that that is something that has been a real lesson for me in — okay, people are not off the top of their heads just saying abortion is a driving thing for them. However, if you make it high salience, you give them personal stories, they react very strongly to it. And so you can move people quickly on that issue, but you have to keep the salience high.

SARAH LONGWELL:

And so back in 2022, there was a lot of men that I was talking to — there was a lot of different operatives who were like, “I think we’re overdoing it on abortion. I just, I think it’s too much. I think we’re talking about too much.” And I was like, “Nope, nope, nope, nope, nope. I got, I promise you, you got to keep this high salience. It has to be top of mind. It can drive people’s emotions. It can drive their vote, but only if you keep it high salience, you let it fall down, people will talk about the economy, they will talk about inflation, they will talk about immigration because that’s where people go much more naturally. They’ll talk about crime.” And so I, I just, I think there’s lessons in there that would be difficult to, how do you, how do you replicate then where — because we do live in community and form our opinions in community. And so in some way, a focus group can be artificial at times. In other ways, it can be more reflective of how people in the real world operate, which is that they bounce off each other in certain ways that change how much they care about something in a moment.

ARIA:

You know, you’re talking a lot about sentiment analysis and you know, until now you’ve only done that with words. And you, what, what is that? There’s that, I don’t know, I think this is true, that there was someone who could watch videos of married couples with the sound off and find out whether they were going to make it or not. Because it’s all about the body language or it’s all, you know, so much of the sentiment of a person isn’t in the written word. And so you, like you said, “Oh, you know what I, I’ve, I’ve learned, I can watch, I can see, I can hear their tone of voice.” You can train an AI to do that too. So it’ll be really interesting when we start using an AI on video as opposed to just words that like, they will get — they — the AI will get so good at the sentiment analysis.

ARIA:

And I think one of the things that could be exciting or terrifying is using the AI for persuasion. It’s like, once we know what works to persuade, is there a way that we can have, you know, a chat bot or an agent talking to someone? Is there a way we can use that to our advantage? And again, that sort of like, sounds dystopian and scary. We’d want to be making sure we’re using it for the right thing. But just like an educational AI can teach young people, can we teach it to have people be less sexist and less racist? And, you know, sort of all of these things for good. Like there’s values there, but it also, you know, there’s some, there’s some qualms.

SARAH LONGWELL:

Yeah. I mean, the way that we’ve used the focus groups strategically, one of the main things that we learned from them early on as a persuasion tool was the importance of messengers. Who was the person delivering the message and how the messenger was oftentimes more important than the message. There’s a bunch of key things that we learned going through the 2020 cycle, and one of them, the reason — so I run a group called Republican Voters against Trump. And the strategy that we employ is to have real people who identify as Republicans, lifelong Republicans, many of them Trump voters. And they say, I voted for Trump the last two elections, but because of January 6th or because of abortion, I’m not going to vote for him again. And, and, you know, they explain it. And oftentimes the videos that perform the absolute best are people who say, “I’ve never voted for a Democrat before in my life, or, I don’t like Democrats or I don’t like Kamala Harris, but I’m going to vote for her anyway because Donald Trump is such a threat, or I hate him.”

SARAH LONGWELL:

So, you know, whatever, whatever language they use, we do know — we are never prescriptive. But we collect these video testimonials and we turn them into advertising. Now they are, they test at the top of the persuasion tool. And the reason is because the number one thing — and this is where I want to sort of get into maybe what I, what I would worry about is — it’s the authenticity of the messenger. It’s the authenticity of, that’s true. And when we, people criticize us, when people are looking to sort of diminish our program or a project, they’ll say, “Well, these are paid actors,” or now we’re starting to hear, “These are AI.” And it’s interesting because one of the things that you’re dealing with, and this comes from having listened to voters, one of the things that’s just happening to people is the absolute collapse of trust.

SARAH LONGWELL:

Trust in institutions, trust in, I mean, politicians were never super high, but like everybody that people used to trust, like has gone, has dropped in the quotient of trustworthiness. And the number one still trusted group of people is people like me, not me, Sarah Longwell, but people like them. We work very hard now to test as a persuasion mechanism like John from, you know, Tommy from Texas is a Christian and it, you know, whatever. And we’re like, “Who does Tommy persuade the most? Like, who does he work best with?” And then we target those people. I can see AI helping us get a lot better and more granular than we already are in terms of which person you were like, well these, this woman works the best with these people. And like you could just tell somebody like, we learned all kinds of things. Like somebody with an accent always codes more Republican for voters, they view them as more authentic. And so the authenticity of the messenger for us is key. And so I wonder what the world will look like when AI is so prevalent that people no longer trust at all the idea that these are the trusted messengers that they want to be persuaded by. My challenge, I guess is I wonder, knowing that authenticity is the key to the realm here, especially in politics, what do you do with AI around that?

REID:

You know, one of the things that I also thought that I love about your particular story is, you know, you’re smart, dedicated, exploratory to technology. But one of the parts of principle, which you know, is too rare in the political environment is this notion of, “I’m going to try to figure out what is the right thing and stick to it.” Like you are a, you know, kind of die-hard, you know, fight-the-battle Republican on deficits and other kinds of things that actually, frankly, I agree with, right? And and, and you get confronted where your pack kind of goes off the deep end and you’re like, “Wait,” right? Like, it, like truth telling is important. Caring about other people’s important. Rule of law is important. Not lying through your teeth and being completely narcissistic is important. You know, you kind of ended up suddenly kind of in this very isolated direction, you know, how did you navigate that? What was the way that you found the kind of personal fortitude to do that? Say a little bit about that journey for you.

SARAH LONGWELL:

Yeah, I mean, the honest answer is that, I think it’s because I had — like, I grew up in a small town in central Pennsylvania. I, my parents were Republicans and I grew up in a very sort of just conservative community. And then when I was in college, I went to Kenyon College in Ohio, which was a great place and kind of just a standard liberally liberal arts school, but had a pretty conservative political science department. It was a Straussian, kind of people from University of Chicago, that I was —

REID:

Oh, Straussians can’t say that they’re Straussian.

SARAH LONGWELL:

That’s true, that’s true. I guess I, I mean, it’s just as a descriptor of the, of the overall program. And the thing that happened is I went to work for a conservative think tank right out of college. This is 2004, 2005. And so Massachusetts is about to become the first state to legalize gay marriage. And I was trying to come out to my parents, and I was trying to come out. And so I was already kind of in the conservative movement had done you know, like had a job in the conservative movement was, was on the road with Rick Santorum and gay marriage was like the topic. Rick Santorum was famous actually at the time for comparing bestiality and homosexuality. And it was deeply uncomfortable for me. I was trying to figure that out. And I was young and I wanted to do a good job, and I didn’t want to reflect badly on my employer or on Santorum.

SARAH LONGWELL:

But also I was trying to figure out how to like come out and be, be out and, you know, I basically had to do it, like I had to, I had to, I — there was actually this moment where there was a, at the, the gay protesters would come to the Santorum events and there were these two moms there. And I would always have to — because I was like the young comms person, the junior comms person — I would go stand with the gay protestors. And so there’s two moms and they’ve got their like tween daughter there. And she’s holding up a sign that says, “My two moms take me bowling.” And I was like, “I quit. Like, I’m out of here. This is, I can’t do it.” It was, it was, and it was this thing that at the time I actually made a promise to myself which is one, I would do something.

SARAH LONGWELL:

I would be out, I would always be out. I would never lie about who I was. That was one promise I made to myself. And then the other was like, I would do something affirmative for gay rights. And so when I moved to Washington, I joined the Log Cabin Republicans, and this is when I learned what it felt like to be hated by everybody. Because Democrats didn’t like you because you were a Republican and said, “Gay Republican, you’re an oxymoron. It’s like jumbo shrimp.” And Republicans didn’t like you because you were gay and you weren’t accepted. And I just had to figure out how to be fine with everybody being mad at me [laugh] or not really having a place in either tribe. And I think that that experience, and I spent years then working on Don’t Ask Don’t Tell repeal, on marriage equality.

SARAH LONGWELL:

I worked on Young Conservatives for the Freedom to Marry. We did a lot of work on marriage equality among Republicans and trying to get Republicans elected who were pro-marriage equality, doing the gay marriage stuff, the marriage equality work, also showed me that people can be persuaded. I mean, the, if you just think about the fact that when I got involved in Log Cabin like 2006 or something like that — between 2006 and 2013, gay marriage happened. Like all the states passed, like 2008 was all the referendums. But I watched it be an enormous issue, a defining issue, something that separated people. And there were just a bunch of groups that tried to figure out, how do you persuade people? And you just saw everybody’s mind change in real time. And by 2013, which was the year I got married to my wife, the Supreme Court, that was the year that they decided Windsor.

SARAH LONGWELL:

And then there was another, but like, it was happening, it was over, the Supreme Court was ruling, like we were, we were done. And then just recently, you know, it passed legislatively. Like it’s, and that is because people change their minds. And so I believe in persuasion. I believe that if you find the right ways to listen to and talk to people — and I spent so much time during that conversation around, “Okay, how do Christians think about gay marriage? How do Black communities and Black evangelical communities think about gay marriage?” You know, like thinking about all these different groups and how trying to understand them, trying to figure out how you talk to them, trying to figure out what it meant to be a human, talking to another human. Like because coming out and love were ultimately kind of the two things that really mattered.

SARAH LONGWELL:

Knowing somebody who was gay that you cared about, and talking to people not about rights, that is the lesson that formed me and shaped me. And so when Trump came along, this was just another instance where I said, Donald Trump is a bad person. Like, that was evident from the start. And sometimes people are like, “Well, it must have been hard for you, you know, you gave up your community or you did this or that.” And I was like, you know how sometimes there’s just like, you can’t do anything else but that thing? Like, I wouldn’t have known how to do anything but that because I wouldn’t have been able to tolerate it. For me, the hardest part was just finding the tribe that like me wanted to do something about it, and like huddling up with that group and being like, “All right, now let’s go. We’re all doing this.”

ARIA:

I mean, I’m going to cry. I have chills. It’s just amazing. And it’s so good to remember because especially in these times when we’re like, “Well, it’s 50/50, we’re polarized. We’ve always been 50/50.” It’s like, not only did I, Pete Buttigieg said you know, recently, like the transition that he’s seen in his lifetime is unbelievable. And also the coalitions have changed enormously. It might be 50/50 today and 50/50 30 years ago, but it’s not the same 50/50. Republicans and Democrats have taken different, you know, points of view and a lot of positions have changed. And so I think giving people that hope that persuasion works, like we can move the needle in the good and bad direction, you know, both happen. And so one of the things that everyone is worried about when they think about this 50/50, how do you persuade is of course, misinformation. And that’s both because of new technologies, but also because misinformation, you know, has been around as long as time. And so I’m sure that in your focus groups, you, you hear this like, how does misinformation — what is the stuff that’s sticky? Like when are you like, “Oh, yep, this, everyone this week is bubbling up on this because it’s stickier, because X.” And what is the stuff that people can be immune to? I’m sure you have lots of thoughts and like, maybe how we can fight against it.

SARAH LONGWELL:

I mean, the biggest example of this is when Donald Trump said that the election was stolen. So the way that this — the information works and how tribal it is, has, it was — I’m, I’m glad, very glad to have had the experience of living through some big lies. Not to not yes, the big lie — but also like big lies to watch how voters have responded to them over time. Because one of the things that was so interesting to me about the election denialism is that a bunch of Republicans right after the election, they didn’t think it was stolen. Like that’s not something you heard from everybody. Donald Trump was saying it. But there’s this phenomenon I call, it’s kind of the “Republican triangle of doom.” And it is the toxic and symbiotic relationship between the right wing infotainment media, the elected officials, and the voters.

SARAH LONGWELL:

Okay? And they work on each other. And when Donald Trump, right, is, is, is just constantly — and, and many of his, the, the elected officials, right? They’re saying the election was stolen, right? Some of them, the most extreme ones, and Donald Trump are saying it — well then there’s a certain amount of media that picks up on that, right? And then there’s a certain appetite from a group of voters that wants to hear that information and they seek it out. And that’s where if Fox News isn’t saying it, Newsmax is going to say it, or OANN is going to say it. Okay? So then they start saying it, and the voters who want to hear it gravitate to those media outlets. Well, then the politicians who want to be on those outlets you know, they, they go on and then they say, the election was stolen.

SARAH LONGWELL:

Pretty soon you’ve gone from, this is a Matt Gaetz, Donald Trump thing to Kevin McCarthy is saying that the election was stolen, right? Like people that you take seriously, and then the more serious the person, well then a whole other set of voters who didn’t think the election was stolen says, “Well, man, if Kevin McCarthy says that the election was stolen well, maybe now I believe it, too.” And pretty soon you have 70% of the Republican party who believes that the election was stolen. And they all work together to make that happen. And Fox News has to compete with those other places. So they say the election was stolen even though they know it’s wrong, even though they know it’s a lie. And so you, you know, I — how you combat it, the only real way is with the truth. And it’s, it’s a tough thing to say because you can’t harness information.

SARAH LONGWELL:

And like the human soul is the, if it wants to lie, like if people are going to lie like, the best way to do it is for us to continue to tell people that lying is wrong. And for us to believe that lying is wrong. Like you do have to try to persuade people. And here’s another thing. So here’s a, here’s a good example of something where I think about, how do we recalibrate the way voters are thinking about things that are essential. So right now, voters want an outcome. They want, they want, they want some kind of policy outcome. They do. They want better healthcare. They want you know, they want lower tax, whatever. But they voters also do not want politicians to compromise, right? The mechanism by which policy happens, which is compromise between the parties, voters no longer have patience for, they’re no longer interested in it. 

SARAH LONGWELL:

They want you to fight. They are mad about the uniparty. They do not want any amount. And so you see politicians now — used to be that politicians would say, “I’m going to reach across the aisle. I know how to get things done. I know how to work with people.” Yeah. But now everybody’s, they say, “No, I, I won’t do anything. I’ll never work with Democrats.” Well, here’s the thing. If you no longer believe in the mechanism of compromise to get to a policy outcome, what do you do? You reach for a strong man. You reach for somebody who says, I alone can fix it. When I think about what is the number one thing that we could do to help with politics, it’s that to reteach people that compromise is actually a good thing.

SARAH LONGWELL:

That it is how they get the things that they want out of policy and politics, which is why politicians go there, is to figure out how to solve problems for Americans. Right? Well, how like, to me, like that’s the answer. The answer is you’ve got to, you’ve got to go back to basics with people because we’re in this enormously disruptive moment. And because people don’t know what to trust, you’ve got to like, rebuild the foundational building blocks. because those things have been eroded. And we’ve got to build them for a new moment with a new information landscape.

ARIA:

I mean, it’s so interesting. I was listening, I think, to Nancy Pelosi on Ezra Klein, and she was actually, she was talking about Don’t Ask, Don’t Tell. And she was talking about how, you know, she had said to some folks, “Well, you’re going to vote and you’re going to make sure that we can, you know, repeal this bill.” And they were like, “No, I’m not. And she’s like, well, you are.” And then she’s like, “And then I went to some Dems and I said, well, you’re going to vote for, you know, military funding to be upped.” And they’re like, “No, we’re not.” “Yes, you are. You are going to do that.” And some people could look at that and say like, “Oh, how cynical how — ” And it’s like, well, no, actually it’s positive. It’s compromise. It’s how we move things forward. And so some of these sort of, the brilliance of reaching across the aisle has been lost. And we have to, we have to get back there.

REID:

Yeah, I think one of the really key things, how do we talk to each other? How do we get compromise? It’s actually, in fact, bipartisan compromise to get shit to happen, right? Like let, let’s get the, the Sinema-Lankford bill through rather than have Trump kneecap it because he wants to run on it as an election. Let’s have it in place for a year helping Americans. You know, that kind of stuff is, I think the compromise and, and kind of work together should be the top virtue.

ARIA:

So Sarah, you said, you know, you have been a lifelong Republican and now you are someone who is trying to get Kamala Harris elected. Where do we go from here? What happens next for the Republican party or for politics more generally?

SARAH LONGWELL:

Yeah, look, I think one of the things I try to explain to people — and oftentimes I am the one person on the panel that nobody’s ever heard of. And I’m like, “Look, if, if you remember one thing that the girl you’d never heard of said on the panel, have it be this: That the Republican party is not going back, there is no — ” Because people ask me this all the time. Like, well, “How do we get back to the Republican party that I knew as a, as a younger person, the one of Mitt Romney, the one of Liz Cheney.” And I just think it’s really important that people don’t spend their time thinking that the Republican Party is going to return to a thing it once was. And the reason that it’s not is that the voters don’t want it. And what’s important, like right now?

SARAH LONGWELL:

Donald Trump has actually been in our lives for almost a decade as a political figure. He has obviously much — been there much longer than that. And so people who have come to the Republican party in the last 10 years have done it because of Donald Trump. They’re not staying in spite of him. They’re there because of him. And so, especially young Republicans today, they don’t even understand why Nikki Haley would be a thing. They’re like that, she doesn’t believe what Republicans believe, what Republicans believe are. We’re isolationists and we believe in tariffs, and we’re America first, and we don’t want any immigrants here. That is not what the Republican Party used to stand for. But a Republican in the mold of the Republicans that I liked when I sort of became a Republican those people have, they’re all gone. They’re gone. And maybe you could still sneak one through here or there in a state like Maryland with a Larry Hogan or something like that, but mostly not.

SARAH LONGWELL:

And so I think it’s important to understand that our politics going forward, the Republican party, you know, is going to look a lot more like some combination of Tucker Carlson, JD Vance, and Candace Owens [laugh]. 

 

ARIA:

That is a terrifying image.

 

SARAH LONGWELL:

I’m sorry. I know, I know. Now I’ll say I do think there is like a vaguely, there’s a healthier version of the Republican Party that is not one that appeals to me, but is one that I think there is a vision among some Republicans that they become sort of a multiracial working class party where you sort of have this political realignment that’s happening and college educated suburban voters, they tend to sort of politically realign into the Democratic party. And the Republican party becomes much more working class and much more sort of oriented around populist economic issues, tariffs, you know, much more isolationist, both in terms of wars and in terms of how we do you know, how, how we sort of trade in the world. And that as long as they abide by the rule of law and election results, doesn’t have to be a dangerous party.

SARAH LONGWELL:

I don’t think it’s going to look, from an issue set, like the old Republican party, but it doesn’t have to be dangerous. And I think that we should endeavor to — wherever it’s going — have the Republican party at least not be dangerous. Which, sometimes people ask me how you do that? And, and for me it really is through sustained electoral defeats because you have to change the incentive structure. And so Donald Trump’s election lies have to have consequences — electoral ones. And that adopting these things have to be sort of politically unviable which is why I find myself working not just to defeat Trump, but candidates like him.

REID:

So, you know, given we’re, you know, it’s kind of possible the tech thing, I kind of want to go to a tech thing for our finale before we get to rapid fire. It’s a little bit long, but I wanted to get to this, which is the, you know, kind of the ability for AI to help guard against misinformation. So we gave GPT-4o, three social media stories we pulled from Snopes. Two are misinformation and one is real. The results looked promising. But let’s quickly summarize GPT-4o’s responses, and can you let us know, could AI actually help combat misinformation or is this wishful thinking? So I’m going to go into the — it’s almost like the, you know, two truth, two lies and a truth, you know, kind of thing. And then, and then have you react. So here are the stories.

PI:

NASA issues an alert for an asteroid that has a 72% chance of hitting earth in 2038.

PI:

Schools stopped teaching students how to write in cursive because AI tools cannot read it.

PI:

The US trade deficit with China has reached its lowest level since 2010.

REID:

So first GPT-4o correctly identified the first two headlines as false and was also pretty good at providing a justification for that. You know, it was for the first headline, GPT-4o said, “NASA and other space agencies closely monitor potential asteroid threats and such high probability of impact would be widely reported.” 

SARAH LONGWELL:

Big news. Feels like big news. Yeah.

REID:

Yes, “and addressed with significant public awareness campaigns.” And when the last headline required more verification, it offered to dive into current data and provide citations from the Bureau of Economic Analysis to verify the facts as true. So what do you think — could this have more widespread applications for debunking misinformation? And, you know, what might we do with GPT-4o to make its debunking more convincing for people who need convincing the most? And, you know, part of this is the focus groups, and does it feel like it’s talking to me and is it trustworthy and that kind of thing?

SARAH LONGWELL:

Yeah, I mean, a lot of this is the way that when something like this rolls out there will be an effort among people who want to tell lies. Sorry. So Donald Trump, one of the, one of the most interesting things I ever saw him do was when he was talking to Lesley Stahl and she asked him why he beats up on the media so often, why he talks the way he does, calls them enemy of the people. He said, “Well, I don’t, I want — so that when you guys say something bad about me, people don’t believe it,” right? And so people who want to be able to lie with impunity will do anything they can to create doubt on any tool that would hold them accountable, right? And Donald Trump does this with everything. So if the courts are holding him accountable, they sow doubt with the courts.

SARAH LONGWELL:

If it’s the police, then the police are bad, you know, the ones you know, because like — so I think that the question is, “How do you roll it out in such a way that, that gives it widespread confidence?” Because every time it gets something wrong or identifies, and, and this is like a good example, this is like a human one, but when Meta and Twitter both decided not to, they throttled the New York Post story about the Hunter Biden laptop, you know, that on the right was one of those things where it gave a ton of credence to the idea that these social media platforms were censoring real information. And so you sort of can’t mess it up — like the, like the, the, the proof is going to be, does it do it at an accurate enough way that people come to trust it.

SARAH LONGWELL:

Now it could end up being like a Wikipedia thing where everybody knows this is like mostly good and it’s good enough for me getting the gist about something and I can quickly discern whether something is mostly true or mostly false, but like, it’s not a hundred percent accurate. You know, it could be like that. I guess I’m not, I’m not sure about its application. I will say it has always seemed to me that when it comes to AI, the only way to police bad uses of AI is going to be with good uses of AI. Right? Like if it exists, and it’s going to be used by people who have malevolent intentions that like if something is an AI generated piece of misinformation or meant to you know, lie to people that AI is going to be the tool that somebody else is going to use to prove that it’s false and like, you’re going to need that. I just, I, I hope that it is good enough and accessible enough to people so that it’s not just the malevolent actors who have the most access to it, right?

REID:

I think you’re very wise. I mean, this is the thing that most technology critics miss, which is say, “We’ll just stop the technology.” It’s. “No, no, no. You have to shape it to be in the positive outcomes.” And so the question is, “What do we do to have AI be helpful here in bridging this understanding of truth and compromise? And what are the things we need to do to instill that trust? How is it built? How is it done?” And, and people who disagree with it. And I think you’re completely right, that people tend to attack what are otherwise frequently good institutions or reasonable institutions, whether, whether it’s the courts or police or everything else, because they’re like, “I don’t want to be held accountable.” Institutional renovation is the ethical thing versus burn-it-down. We will move to rapid fire. Very well said, Sarah. Is there a movie, song, or book that fills you with optimism for the future?

SARAH LONGWELL:

So, I will tell you, when it comes to music, I often haven’t been a big music person, but I endeavored during this political — I listened to nothing but political podcasts in 2020, and I found myself on the verge of a nervous breakdown. Like, and I was like, “So, I’m going to get into music going into 2024.” And I’ve been listening to this young guy named Zach Bryan. I’m a kind of a country music person, but Zach Bryan is this 28-year-old who’s already put out like seven albums. Every single one of them is a banger. I listen to all of the songs on shuffle and it makes me so happy because it speaks to something universal in the human condition, which for me, when I think about optimism about the future, it is like, how are we going to relearn how to connect with each other? And not in like a silly way, like, “Well, I’m going to figure out how to set us all at a table and talk to each other.” Like, we just have to remember that like, we are so mired in our tribalism and we have got to keep our eye on the bigger tribe.

ARIA:

I love that. Although I have to admit, I thought you might say JD Vance’s “I’m A Never Trump Guy” remix.

SARAH LONGWELL:

Oh my God. That is so funny.

ARIA:

That is one of my favorite [laugh]. Every time I see that dance, I just, I smile. Okay. Rapid-fire number two. Is there a question that you wish people would ask you more often?

SARAH LONGWELL:

My favorite thing is when people ask me about the people I work with at The Bulwark. So one of the great joys of my life now is that Tim Miller, who I work with at The Bulwark, and Jonathan Last. Tim Miller’s a very old friend of mine and I met Bill Kristol and Mona Charen and a lot of the other people who worked there during this moment. And telling the story about how we all came together and what it’s like to work together like it, I love talking — they’re the smartest, most interesting people, and I love talking about The Bulwark. And so anybody who asks me something about The Bulwark. Always brings me joy.

REID:

So where do you see progress or momentum outside of your industry that inspires you?

SARAH LONGWELL:

So I’ll go back to the fact that I’m gay. And one of the things that — and, and specifically a lesbian. So one of the trends that I have found a little troubling has been as everybody’s moved into queer, as like a catchall, that there was like a little bit of erasure of just lesbian stuff. You know, we like Subarus and softball and being a lesbian. I remember being, I, when I sort of moved into lesbian culture, how wonderful it was to like find a group of people. It was like, “We got jokes, guys. We got jokes. Stereotypes are true. They’re true for a reason.” And so for a while I sort of felt like the kids weren’t really that into that. Like we were like old fuddy-duddies to use the word lesbian and talk about Ani DiFranco.

SARAH LONGWELL:

But like, there’s a whole new world now of Billie Eilish and a bunch of like, there’s a song like it’s very much in pop, it’s like reemerged in pop culture in a way that I am finding like the kids are going to be all right. I, and, and, and I, I say that a little bit jokingly, but I will say that like the trend going back to like the way that gay families exist now in the world and the way that we see ourselves reflected in a lot of the culture has just been — even with the backlash against “woke” and like the way that Ron DeSantis kind of brought back, like, “Don’t Say Gay”, and like it felt like we were backsliding. I actually still feel like the world is on the side of people getting to be who they are, love who they love, and like is very much embracing the new way that families look, like Pete Buttigieg talks about. And that is a trend that — just like a life trend — that I feel happy about all the time, to the point where when I rethink about it, it makes me happy.

ARIA:

Can you leave us with a final thought on what you think is possible to achieve if everything breaks humanity’s way in the next 15 years, and what’s the first step to get there?

SARAH LONGWELL:

So here’s the thing that I want people — right now in our polarized politics — what I want them to know about people. So I listen to people all day, every day, and people are the best. People are super funny. They volunteer in their communities, they take care of their sick parents. They’re all worried about how their kids are going to do. They’re all worried about the country. And we have so much more in common than we do — you know, it’s like cliche to be like, “What unites us is greater than what divides us.” But it’s also true. And like, if we can just readjust our frame. Like we can head into civil war if we want and decide we hate each other but we definitely don’t have to and actually really don’t want to. One of the things I hear people say in the groups all the time is they’re like, “You know, I’m inside, I’m on my computer and everyone’s so mad, whatever. And I go outside, everyone’s so pleasant and they wave to you and they say hi.” And it’s like that world exists for us. We just have to figure out how to, how to seize it and bring it back. We do not have to be angry all the time.

REID:

I think that is an excellent, excellent way to wrap it up. And Sarah, thank you for joining us on Possible.

SARAH LONGWELL:

Oh, thanks for having me. It was so fun.

REID:

Possible is produced by Wonder Media Network. It’s hosted by Aria Finger and me, Reid Hoffman. Our showrunner is Shaun Young. Possible is produced by Katie Sanders, Edie Allard, Sara Schleede, Adrien Behn, and Paloma Moreno Jiménez. Jenny Kaplan is our executive producer and editor.

ARIA:

Special thanks to Surya Yalamanchili, Saida Sapieva, Ian Alas, Greg Beato, Parth Patil, and Ben Relles. And a big thanks to Tony Franquiz, Priya Gada, Natalie Castonguay, Conor Kilgore, Meaghan Leister, Karrie Huang, and Little Monster Media Company.