Bernard L. Schwartz Lecture With Evan Spiegel of Snap Inc.

 

Snap Inc. Cofounder and CEO Evan Spiegel discusses the growth and evolution of the company, implications of technological innovation on data privacy, developments in artificial intelligence, and how businesses are adapting to geopolitical and macroeconomic uncertainty.

NICHOLS: OK. The eagle has landed, as we say elsewhere. (Laughter.) It’s my great pleasure to welcome you all here to the Council on Foreign Relations meeting with Evan Spiegel, the cofounder and CEO of Snap.

This meeting is CFR’s Bernard L. Schwartz Annual Lecture on Economic Growth and Foreign Policy, which features speakers with expertise at the intersection of business and policy. I think that’s you. Bernard Schwartz was chairman and CEO of Loral Space and Communications, and a Council member for more than twenty years. Since his passing last year, this lecture series continues in his honor and memory.

My name is Hans Nichols. I am a political reporter at Axios. And I will be presiding over the events in this hour-long conversation. Just to remind everyone, I always say as a journalist, and Ambassador Froman, this is an on-the-record meeting, right? I’m supposed to say that. The first thirty minutes you and I are just going to be talking, and then we’ll take questions, the better questions, from the audience, which will be—will be all of you.

You know, I want to start off with AI, because it’s so obviously the topic of news. I was just—you know, Senator Kelly released a new report here, AI for America. You know, there’s billions—I mean, there’s headlines all the times, billion dollars spent on data centers and this and that. But it’s still kind of confounding as a lay person. And you’re covering this, or understand it a lot better. Walk us through the optimistic case and the pessimistic case for AI.

SPIEGEL: All right. We’ll jump right in. (Laughs.) Well, thank you so much for having me. It’s so great to be here. And sorry I’m running late. For the first time in my fourteen-year career I’ve been late to a speaking engagement today. But also for the first time in fourteen years my dad wanted to come to this speaking engagement, of all speaking engagements. (Applause.) So my dad’s actually with us, which is—I got an email. Evan, do you think I could come to this? It’s, like, of course, Dad. It’d be great. So thanks, Dad, for coming.

And now let’s talk about AI. So I think for me, the optimistic case is that we point AI at some of the biggest problems that we have today as a country. I think if you look at the problem—the biggest problem that most Americans are facing today, it’s an affordability crisis. It’s cost of living. And I think we could choose to apply AI to solving those problems, right? Whether it’s things like, you know, some of the pilots that are already happening today to approve, you know, housing plans, right, faster through the city. This is something that’s being trialed in Los Angeles in the wake of the fires, but also in many other places.

I think it certainly could be applied to health care, right, where 15 to 25 percent of health-care costs are administrative, non-care costs. So I think the optimistic case here is that we aim AI at some of these really serious problems that are facing Americans, and in doing so build a lot of trust with Americans that AI is a good thing. I think today if you look at the polling, most Americans are very concerned about AI. They’re worried that AI is going to take their jobs. And so I think if we want to see more progress with artificial intelligence, we’re going to have to build trust with Americans by solving the problems that they face today. So I think that’s the optimistic case.

I think the pessimistic case is that AI gets primarily applied to large sets of high-paying jobs in the United States. That it happens faster than people are expecting. That we’re not able to use tools like AI to upskill the workforce fast enough to offset that transition. That AI is then, you know, integrated into our weapon systems and nuclear weapon systems in ways that don’t have humans in the loop, and we could go on and on from there. And I think it’s important to also recognize that there have been—you know, if the allegations are true—tragic cases already where people have, you know, had their lives seriously impacted, or in some cases even lost their lives, you know, because they’ve formed a relationship with artificial intelligence chatbots.

So I think we need to recognize that there already has been a very real human impact. That I don’t think it’s too late, you know, to fulfill the optimistic case for AI. And also, that if we want to really lead in this domain that building trust with Americans by solving their problems is probably the best way to do it.

NICHOLS: Is there a specific case for you where you think, boy, I’d like to have an AI agent sort of deployed for that?

SPIEGEL: I mean, my job would be—(laughter)—I think for us what we’re finding is that, especially in areas where we have repetitive tasks that require very high judgment, that already we’re seeing AI outperform human agents, and doing that at tremendous scale. So for a platform like ours that reaches, you know, nearly a billion people, about 930 million monthly active users, we want to really create a safe environment on our platform. That means that we moderate content. Doing that at scale with 930 million people would be almost impossible using human agents—probably impossible using human agents. But I think, you know, leveraging AI we’ve been able to, I think, remove a huge volume of otherwise inappropriate content, content that doesn’t meet our guidelines. And, you know, have created a safer and more positive environment for our community. So these tools are already deployed at scale within our business in ways that have had profoundly positive impacts on our community.

NICHOLS: I know you’re Los Angeles based, but further up the coastline and in the Bay Area, they have this concept of P(doom), sort of the probability for doom, and a really sort of negative outcome, right? Elon Musk puts it about 20 percent to have sort of an extinction-level event. Earlier today we—actually, I was at an AI conference. And Amodei said something, like, 25 percent chance that AI goes really, really badly. What’s your sort of probability? What’s your number?

SPIEGEL: I define really, really badly. I think an extinction event is highly unlikely. I’m very optimistic about humanity overall. I think we’ve navigated profound technological changes before, some of which I would put on par, obviously, with a transition like artificial intelligence. So I’m quite confident in humans’ ability to navigate this transition. And I think, fortunately, we have AI to do it as well, to help us. And so I think every time, you know, we’re concerned about the impact of AI, we should also be thinking about how we can use AI to solve that problem as well, right? I think the content example I just gave was a good one. We’ve seen a proliferation, right, in content that’s been created by AI, but at the same time AI tools are better than ever at detecting that content and removing it at scale. So I think we really need to make sure that we’re constantly thinking not only about the probability of doom, but also all the ways that AI can help us solve these problems.

NICHOLS: So above or below 20 percent?

SPIEGEL: I’d say healthily below 20-25 percent.

NICHOLS: Ten percent?

SPIEGEL: Probably below 10 percent. I don’t think there’s going to be an extinction event. Yeah, I don’t—

NICHOLS: OK, no, I don’t—you know, I don’t have strong opinions. Look, my greatest hope for AI, and I’ll get made fun of for this later. This is my quirky thing. Is it will teach us so much about things we don’t know. And so I think there’s always optimism in the quirky thing that I’m embarrassed to say. But we’re going to figure out how dolphins communicate with each other. It’s a large language model. They clearly talk to each other. And, like, a large language model will be able to, like, potentially decode how they’re talking. So mock me later, but that’s my—that’s my quirky take on it.

We’ll move on to another serious aspect of AI, and that is China. And I feel like the conversation, whether you’re talking to a Democratic senator—I mean, you know Senator Kelly, a Democrat from Arizona, was talking about this today. China, Ted Cruz talked about it also today. And it seems as though this sort of—the contest—I don’t want to call it a conflict—but the contest on who’s superior on AI is going to unfold pretty quickly. Do you agree with that sort of timeline? Is it this decade? And how—like, what steps does the tech community and policymakers in Washington—what do they need to take to ensure that the American model of AI and keeping humans in the loop emerges as the more dominant model?

SPIEGEL: Gosh, I think it’s really hard to say, especially because AI is only one dimension of this competition, right? I think there—you know, we are in an unbelievable race in terms of advanced technology with other countries, like China. And I think, you know, it’s way too early to call a winner. I think especially with a country like China, that has so many unbelievably talented people, that has invested so heavily in engineering education. You know, it’s very difficult for a country like ours with, frankly, a smaller population, with less, you know, investment in engineering and advanced technology. I think we’re up against a very, very serious competitor.

So I think way too early to call. I’m not sure we’ll know by the end of the decade. I think, you know, AI is only one piece of the puzzle here. But I think what’s most fascinating to me is the way that America is adapting to no longer being the only game in town, right? Post-World War II we really had the benefit of being the only show in town. And I think we’re now going through this period where we’re learning how to cope with other countries that are, you know, seemingly as powerful, in some dimensions, as the United States of America. So how we navigate that, how we manage that, I think is more important than almost the—you know, each of the individual technological races.

NICHOLS: Does power supply concern you at all?

SPIEGEL: I think that’s actually one of the things that—you know, if I look at the optimistic column. I mean, I’m just so excited to see so much investment going into new forms of energy, you know, generation. I think in some cases we’ve almost been too religious on energy, you know. And I think we need to take a portfolio approach, which seemingly now I think folks are coming around to. I think there’s going to be room for lots of different types of energy generation. But some of the most interesting technology around fusion and whatnot is finally getting more attention and more investment, because of the enormous needs of these data centers. So I think one of the very positive, happy outcomes we could get from all this investment in AI and in data centers are some real breakthroughs on the energy side. And some of the stuff we’ve seen, you know, in the fusion space is just not that far anymore.

NICHOLS: Yeah, more optimism. Well, we started off kind of at the macro and we’ll just sort of dig down into your company here. Your PR team helpfully pointed out, 75 percent—this number kind of, like, surprised me—75 percent of thirteen to thirty-four-year-olds in the United States use Snapchat every day. You obviously see—like, you know, there are obviously safety and privacy concerns. But you see a lot of the conversation, especially as it relates to politics, taking place there. Did you see the rightward shift of young people ahead of the 2024 election?

SPIEGEL: You know, I think what’s really interesting is that young people in particular have been quite consistent in terms of their concerns, right? I think first and foremost being the cost of living, right? The frustration that they’re not able to purchase a home. The feeling that, you know, perhaps their job might—prospects might change due to the advent of AI. I think that those concerns have really animated a lot of the races, not only, you know, the presidential election, but even some of the, you know, more recent results out of New York, for example, where I think the candidates that are successful, especially with young people, are ones who are speaking consistently to those issues.

So I think to me, you know, as I look at young people—and I think now, you know, that there are more folks registered independent than Democrat or Republican, right? So I think, to me, some folks, especially here in D.C., I suppose, are focused on party alignment. What I’m seeing with young people is that they’re voting for folks who are talking about the issues that matter to them. And I think that it’s really, really important, you know, for our leaders, for politicians, to speak to those critical issues.

NICHOLS: But it wasn’t—you didn’t—like, TikTok, it was pretty obvious. You saw, like, TikTok kind of took a rightward bend to a lot of Trump content, MAGA content really took off there. Did you see anything similar for Snap?

SPIEGEL: It’s a good question. I don’t have any, you know, specific stats to share on that. But I do think it’s not a secret that, you know, the Trump campaign was very effective on social media, right, on creating content that was highly engaging.

NICHOLS: But remind me again, they did not advertise on Snap?

SPIEGEL: I’m not sure.

NICHOLS: OK. I brought up TikTok, so I’ll—you know, following the Chekhov’s gun principle here I will mention it. Do you support this forced sale?

SPIEGEL: I think it is very important that we have a consistent framework to handle communication services that are, let’s say, based in countries where we are in a serious strategic competition. I think if you look at Snapchat, in our business, we’re not allowed to offer our service in China. You know, many other technology companies are not allowed to offer their services in China. At a very high level, that creates an uneven playing field. Because if a there is a technology company based in China who offers a communication service like Snapchat, they can use their profits that they’re able to generate in China, where we’re not able to compete, to reinvest in growing outside of China, right, in places like the United States, where we’re forced to compete with those companies.

And so I think just as a matter of fairness we do need to figure out a framework for how, you know, we’re going to manage competition with companies who have privileged access to markets like China that United States companies don’t have. I think this may be one attempt at getting at that issue, although this has more of a national security focus. I think what I would really love to see is a comprehensive framework for addressing that issue, especially because it impacts more industries than just technology.

NICHOLS: I mean, you look at the auto industry, like any industry you can think of, if you have a consumer base that allows you to generate profits and it can allow you to go elsewhere and undercut, that’s a problem, no?

SPIEGEL: Well, I think in the United States we have taken a position on Chinese EVs, right? We’ve made it almost impossible for them to enter the U.S. market, due to very high tariffs. So I do think the United States has chosen, in some areas, to protect U.S. companies. I think the question is, you know, could we develop a broader framework to really, I think, create more stability for U.S. companies, and Chinese companies, where, you know, we say, hey, there are some areas where it makes total sense for us to rely heavily on China and their products, and there are other areas where, you know, maybe not.

NICHOLS: As a cofounder and now CEO of a, you know, incredibly successful, profitable company, does it make you at all queasy that the government can kind of come in and force you to sell?

SPIEGEL: How do you mean? Sorry.

NICHOLS: Well, you created a company. If the government came to you and said, Evan, you got a great company, but you need to sell it. And I’d like you to meet Mr. Ellison here. And he’s got a check for you. (Laughter.)

SPIEGEL: Well, I think, to be clear, what you’re talking about is a law that was passed by Congress, right? Not an arbitrary act of, you know, the president, for example, the executive branch. So I think in the architecture of our government I think our founders considered these sorts of scenarios, right, and designed processes to try to prevent this sort of arbitrary behavior that you just described.

NICHOLS: OK. That’s a thoughtful, reasoned answer. I will have another question that’s not necessarily thoughtful or reasoned. But I don’t mean to sell out my fifteen-year-old daughter. But I did ask my fifteen-year-old daughter ahead of time. So this is—I’m very glad you might—we’ll bring you up here for this part of the conversation. And I said, look, do your friends use it? I’m interviewing the CEO tomorrow night. If you have any questions for me, let me know. If your friends have any questions, let me know. And this is my daughter: Dad, you know, I don’t have Snap. But all my friends do. You should ask him five reasons why I should have Snap, because you don’t let me. (Laughter.) So five reasons why my fifteen-year-old daughter should have Snap.

SPIEGEL: Well, I think first it might be helpful for me to understand your concern so I can—

NICHOLS: Yeah. We generally think—that was the next question from my wife, which is, have you read the Jonathan Haidt book The Anxious Generation. But I think we generally, and I think there’s almost a bipartisan consensus that you’re starting to see emerge on phones and social media, that sometimes they’re not great for people in their developmental phases. And so, you know, like there are school bans—cellphone bans in states all across the country, that’s like—and different school districts. So our concern, and I guess I’ll speak candidly here, is that there’s too much exposure to social media, and that there isn’t enough in real life conversation. And so that’s the kind of decision we’ve made. But I’m willing to be, and she’s certainly willing—(laughter)—to have you convince me otherwise.

SPIEGEL: Yeah. I mean, I think, you know, rather than make the case for smartphones in general, I think one of the things that people really love about Snapchat is it allows them to stay in touch with their friends and family when they’re apart. And that’s consistently what we hear from our community. It’s consistently what we see from the research as well, independent research. There was great research out of the Netherlands, out of Australia as well, that actually showed that Snapchat—you know, when compared to TikTok, or compared to YouTube, or Instagram—was actually supportive of people’s wellbeing and their relationships. Because it’s a really easy and fun way to stay in touch with your friends. So I think if you look, you know, net-net at the overall health impacts and wellbeing impacts of having friendships and being able to develop those friendships, I think they have a hugely positive impact on mental health.

I think there’s a lot of research that shows as well that even, you know, casual friendships, somebody that just checks in with you, you know, once a day or something like that, can be the number-one predictor of being able to navigate a depressive event. So I think overall, our view is that friendship and those relationships are incredibly important to people’s wellbeing. That Snapchat supports those relationships. And that we also have a lot of safeguards in place, right? So, for example, unlike iMessage or something like that, two people need to add each other on Snapchat to be able to communicate. Unlike iMessage, if you’re in a conversation and you feel uncomfortable, you can just press and hold, report that message, and get help from our trust and safety team. So I think we offer a lot of tools that other communication services don’t that make it easier to have fun and safe conversations with friends. So I think if your daughter wants to use Snapchat to communicate with her friends, especially her friends from school or real life, I think that’s something that that I, at least, would consider.

NICHOLS: OK, that’s one. You owe me four more. No, we can—we can move on. Have you read The Anxious Generation?

SPIEGEL: I’ve read some of it, but not all of it.

NICHOLS: OK. Do you agree with his general assessment?

SPIEGEL: What do you see as his general assessment?

NICHOLS: That kids spend too much time with their phones and not enough time outside.

SPIEGEL: You know, I think it really depends, you know, on the kid. I think, you know, that the general observation you made that we should all have balance, right, is really important. I think the average adult spends seven hours a day looking at a screen. And I think we should imagine and build towards a world where we use computing differently. I think for us, in our household, we just think it’s important to—you know, our fourteen-year-old, for example, uses Snapchat, but he also really likes to hang out with his friends. In fact, he would prefer to hang out with his friends in person. But if he’s not able to see them in person or, you know, for one reason or another, allowing him to communicate with his friends on his phone is something that we’re OK with as a family.

NICHOLS: Walk me through the limits. I read in People magazine that has a firm 9:30, no more phone.

SPIEGEL: 9:30 phone and computer outside the bedroom.

NICHOLS: Outside the bedroom. And then what time can he use it in the morning?

SPIEGEL: You know, that’s never been an issue because we’re usually waking him up, so. (Laughter.)

NICHOLS: For your other children, when will you allow them to have their first digital device, and what will that device be?

SPIEGEL: I think probably, you know, we’ll consider phones in middle school, if that’s still—if that’s still a thing. That’s probably something that my wife and I—you know, we would talk about, probably align on that. You know, we try to be thoughtful about Flynn, because, you know, he was spending time between our house and his dad’s house. And so having a phone was actually really helpful and important for coordinating all that. So we were open, you know, to him having a phone earlier. I grew up—my parents didn’t allow me to watch television at all, which I thought was, you know, in the long run, great. In the short term I had no idea what the hell South Park was, which created some social challenges. (Laughter.)

But—(laughs)—ultimately, my parents really encouraged me to read a book, or go play outside, or to build stuff. And I think, you know, to me it’s so much more about supporting, you know, kids’ passions—giving them the opportunity to play sports or pursue something else they’re interested, read a book. You know, I think we’ve become a bit—maybe, in the wake of the pandemic, right? You know, I think back, you know, to the pandemic when, you know, parents, educators, people were, you know, essentially telling everyone to use their phones and computers all the time, right? And now it feels like we’re in a moment where people are telling folks not to use them at all. And I think, you know, we have to find some healthy medium, where folks are engaged in the extracurriculars that matter to them and also use their phone to talk to their friends, you know, when appropriate.

NICHOLS: You have a future as a television host because you made the transition, the segue, that I want to make. Which is, what comes after cellphones? And you hinted at this on, like, for your children when they’re in their middle school. We had a chance to kind of talk about the fifth-generation Spectacles. But, like, what excites you the most about Spectacles, about what you’re doing at Snap? Like, tell us what the future looks like.

SPIEGEL: Yeah. So we spent the last eleven years and more than $3 billion trying to reinvent the computer. Which sounded a little crazy, I think, when we got started, just as, like, a photo-sharing service. But we really very quickly in building our service bumped up against the limits of the smartphone. And so, you know, we invented things like vertical video. At the time all the videos were horizontal from television. We were, like, that makes no sense. You got to turn your phone sideways or watch this tiny video. We invented vertical video. We invented things, like, you know, tap to take a photo, hold for a video, instead of having to toggle, you know, your video settings. Of course, things like stories that put your day in chronological order.

But all of this felt very limited by this tiny screen where we spend, you know, seven hours a day hunched over this tiny glowing rectangle. And we just felt like, you know, computing could be so much better. It could be so much different. And, you know, the vision that we always had that I would talk about with our friends is, you know, I talked about wanting to look out our window one day, see our boys—we got four boys at home—running around outside, turn to my wife and say, honey, they won’t get off the darn computer. Because we believed that there was a way to build a computer that actually brings people together, that brings all those benefits of computing into the real world, in a way that can be shared.

And so we’ve built that with Spectacles. They’re see-through glasses. They’ve got, you know, basically a supercomputer in a pair of see-through glasses. That allows you to put computing in the world, share it together with your friends so you can build things together. Our kids run around and play outside together with the glasses. And I think this is the beginning of a totally new way of using computing. And it’s obviously being accelerated by AI because we’re going to spend less time hunched over operating a computer and more time supervising AI agents that are working on our behalf. And so I think this form factor shift is happening at a really interesting time.

NICHOLS: What’s the biggest technical challenge there? And not—I mean, you know, I’m not an engineer, but it seems like battery life comes to mind. Like, what are you—what are you working through?

SPIEGEL: I think packing all of that performance into such a lightweight pair of glasses is incredibly challenging. Yeah.

NICHOLS: And you’ve sort of taken this approach where you’re—I believe that, you know, creators and developers have access already just some of the, I don’t want to say—you know, some of the earlier models. 2026 is the goal to have a consumer-facing Spectacle. Is that still on track? And is that something that’s going to be available for purchase? Or what does that look like?

SPIEGEL: That’s on track, coming next year, available for purchase. In the meantime, if you want to check out what developers are building you can just go to the Spectacles Reddit. It’s, like, Reddit.com/r/Spectacles. There’s a whole community of people building totally wild stuff. So if you want to just get a feel for what’s coming, it’s worth checking out.

NICHOLS: But is this the first time you’re doing hardware?

SPIEGEL: It’s not, actually. So this is our—what, it will be our sixth-generation—

NICHOLS: Oh, but the Spectacles themselves. This is a hardware product that you’re going to be selling to consumers?

SPIEGEL: Yeah. We sold earlier—so about ten years ago we created camera glasses, kind of like the Meta Ray-Ban glasses are today. And, you know, very quickly, obviously, realized that your phone is quite a bit better at taking photos than a pair of glasses. And very quickly moved—you know, first, you know, adding a second camera to get depth, adding a display to be able to overlay computing on the world, and then working to get that into a form factor that was really powerful, but also very wearable. So that’s coming next year.

NICHOLS: OK. It’s this portion of the conversation that I open things up to the audience. And if anyone has a question. The lady here right front was the first one that caught my eye. And I think you’re supposed to say your name, and then I’ll gather other people. And then fire away.

Q: Hi. Good evening. Thank you. My name is Simone Williams. Millennial. So I’ve seen the ranges of social media, right?

Two-part question for you is, what do you actually think is, like, the next innovative thing for social media? Like I said, Millennial here. So I’ve seen it—like, the different tranches of it. And it almost feels like social media is just taking bits and pieces from individuals. I mean, like, you got Instagram take—in my opinion—taking your stories, and things like that. So what do you think is, like, the actual next innovative thing for social media? And then, taking it back to the earlier conversation on civic engagement, how can social media better help civic engagement? Knowing that individuals stay in their echo chambers?

SPIEGEL: Yeah, what a great question. So big picture what am I most excited about with social networking or social media? You know, with the glasses that we’re building with Specs, we believe that, like, the next social network is actually in person, right? Because if you can actually use computing together in person, if you can see—you know, look through a pair of glasses, and, you know, I think in schools, for example, do chemical reactions together, right? Or, you know, learn about the human body. Or a lot of people like, you know, playing—there’s a game called Sightcraft, where you can throw little, you know, beams of light, essentially, at targets, and run around and do that together with your friends.

I think, you know, it’s unlikely we’re going to get to a place where people just put computers away, right, and sit around a table. Maybe here at CFR. But I think if we—if we make computing something that people can enjoy and do together, that that really, to me, is the most exciting vision of what the social network can look like in the future. I think if you’re asking about smartphone-based social networking, I think you’re going to see a lot—a lot more AI-generated content, generated, you know, based on people’s preferences, and changing very, very quickly. I think you’ll see a lot more—well, we have something coming in a couple months that will be interesting, and hopefully other people will copy too, because it’s a great idea.

NICHOLS: You can go ahead and release that now. It’s fine. (Laughter.) We can put this part off the record.

SPIEGEL: (Laughs.) So, yeah. So, I think there will continue to be some fun things coming in mobile social media as well. But the big picture, I would say, you know, with mobile social media is actually the bifurcation that really started happening maybe back in 2017, where people, you know, are posting less of their personal lives publicly, and instead just watching entertaining content from creators, and then messaging with their friends. And if you really look at the architecture of Snapchat, from the beginning, you know, you sort of have chat on one side of the app, the camera in the middle, and then you have content, you know, from your friends, but also primarily now from creators, on the other side. So I think that bifurcation, the notion of social media that people are going to post publicly about everything going on with their lives and comment on it and like it, I think that’s been decaying for quite some time. And instead, today you see a combination of messaging and then, you know, essentially personalized entertaining content.

NICHOLS: I think Ambassador Froman had a question, so I will call on him. Because he’s the president. (Laughter.)

Q: Thank you, Hans. Thank you, Evan, for being here.

What do technologists, the technology community, not get about Washington? And what does Washington not get about the technology community? I’m asking for a friend who runs a think tank. (Laughter.)

SPIEGEL: I think, if I look at one of the most recent sort of disconnects between Washington and the technology industry, I think sometimes—you know, and maybe this is the misconception on both sides. I think both technologists and folks in D.C., like, these are not just jobs to these folks, right? I think, you know, whether you’re a technologist serving a community, whether you are in politics, serving constituents, I think that framing is incredibly important for both groups understanding each other. And I think—you know, I think that is also where the overlap is most relevant and important.

And so I think, you know, if we can get both, you know, Washington, D.C. and technologists to talk about their constituents and their customers together, and what is best for those customers over the long run, I think it’s going to be easier to get things done. Because I do think both groups are incredibly passionate about what they do and incredibly dedicated to serving their constituents and their community. So that, to me, is always the frame that I think is so important.

NICHOLS: Do they talk past each other sometimes?

SPIEGEL: I’m not sure they talk past each other, but I think—you know, what’s the polite way of putting this? I think both, you know, folks in D.C. and technologists want to be deeply respected for their contributions to our country and to society at large. And I think when either side doesn’t show that sort of respect for one another, that’s when things sort of break down. And that’s what I’ve seen in the past. And I think one of the things that I think has energized the technology—if I’m, you know, being frank—one of the things that’s energized the technology industry recently is that this administration has been very vocal about how critical technology is to winning in the long term and to American supremacy. And so I think, you know, that framing is something that has made the technology community feel valued for their contributions in the U.S., and want to do more to help the United States and, you know, its citizens.

NICHOLS: The gentleman the green pants right behind Ambassador Froman.

Q: Hi, Trooper Sanders.

So January 2029, you are part of a CEO dinner with the president and vice president-elect. We are living in a world where AI has fueled 10 percent economic growth and 10 percent unemployment. What are you putting on the table to address that, that may feel uncomfortable today?

SPIEGEL: You know, the biggest thing I would focus on immediately is the need to build an extraordinary amount of infrastructure here in the United States. I think, you know, we spoke about energy earlier, but I think in almost every corner of the country we have aging infrastructure that needs to be rebuilt. And I actually think one of the huge benefits of AI could be freeing more extraordinary talent to build some of this absolutely critical infrastructure. So I think the primary thing I would be focused on is upskilling. I would be focused on doing that with AI. I mean, one of the things that’s so interesting about glasses, for example, you can just record a video of yourself building something or fixing something, dump it into a video model, and you have a tutorial that literally is overlaid on the world teaching you how to do it, right? So, like, I think there’s going to be so much opportunity to help use these tools to upskill or reskill the workforce. And I think to do so in service of a lot of the infrastructure that really needs to get built here in this country.

NICHOLS: Is 10 percent unemployment realistic?

SPIEGEL: I don’t think it’s unrealistic.

NICHOLS: So you think there’s going to be a serious transition? I mean, you talked—you started at the beginning by talking about the trust that AI and the tech community needs to build with the American public.

SPIEGEL: Yeah. I mean, I think if you have 10 percent GDP growth, 10 percent unemployment is unlikely, right? You know, just given how important the consumer is the economy as a whole. So I think that scenario, you know—

NICHOLS: Ten and ten doesn’t make sense. But is it—if it’s five and ten?

SPIEGEL: Five would still be incredibly high, right, relative to where we’re at today. So I think it’d be unlikely to see 10 percent unemployment with 5 percent GDP growth. I think, you know, the downside scenario is just massive deflation without upskilling and without retraining. I think in general, right, we’ve been very fortunate to be in a tight labor market where the primary concern that American citizens have is cost of living. So I think if you put all the pieces together, this technology is actually coming at a pretty constructive time, where we should be able to leverage this technology to bring down the cost of living and, hopefully, create new opportunities.

I mean, I think if you look at the dynamics of the U.S. workforce in the areas where we’re invested, I think AI is poised to probably make us more competitive, right? If you look at this vision for reshoring American manufacturing, that’s only going to happen with more automation. And so I do think, big picture, the U.S. has more to gain than to lose. But there will be a period of transition.

SPIEGEL: In the back. Yeah.

Q: Hi. Good evening, Evan. My name is Kirsten Asdal. I run a China-focused political risk firm. Thanks for your time this evening.

I’d like to ask about values, and the values through which you see the world. It’s my understanding that the techno-optimist crowd understands technological progress as a universal good, and the introduction of new technology into society—that integration is just foundationally good, objective good. I happen to think that it’s heavily dependent on who is introducing it, what government, what person, what company, and how they’re doing that. What are—what would you say are the values that you live your life by, and that you hope to use Snap to fulfill in the world? This is pretty meta, but it’s—

SPIEGEL: Well, not Meta. Let’s not say Meta. (Laughter.)

NICHOLS: Yeah, I was going to say, actually. (Laughter.)

SPIEGEL: Yeah. Our values at Snap are kind, smart, and creative. That’s, you know, what’s been most important to us. And the reason why we’re so focused on kindness is because it’s so deeply connected to creativity. We’ve just found that unless people feel supported in the workplace, that they can show up and bring wild ideas and be their full self and not be afraid, you know, of failing, it’s incredibly hard to be creative. And in our industry, unless—you know, especially for us, right? We’re a challenger company. We’re a lot smaller than these huge, multi-trillion-dollar companies. Unless we can constantly innovate by creating new things, our company will die.

So I think we’ve really built the company on this foundation of kind, smart, and creative. We didn’t realize what an impact it would have in the early days. You know, Bobby and I, we started the company, we’re, like, what values do you think we should have? Like, oh, you know, we’d love to work with people who are kind, smart, and creative. And then only later did we really, you know, appreciate and understand the deep connection between kindness and creativity. I think that served us well. And it certainly—you know, those are values that are important to me.

NICHOLS: I believe I’m supposed to take some virtual. So if Sydney (sp) can pull them up. Will they be behind me? Or no?

OPERATOR: We’ll take the next question from Esther Dyson.

NICHOLS: Esther, go ahead.

Q: Sorry. Here we go. So you talked about upskilling people. And I’m curious what you think of the notion of instead of upskilling them to do stuff that a(n) AI could do, why not skill them to be kinder? Skill them to be more human? What we really need more of is not more AI. We need more teachers. We need more doulas. We need more healthcare people who actually have time to talk to you and listen to you. That’s a prompt.

SPIEGEL: Well, I think you’re touching on something so important, but now we’re really talking about the role of education in general in the country. And I think your observation really rings true, but also sort of, you know, has maybe inspired me to mention a couple other dimensions, right? We hire folks all the time out of school. We find that they’ve spent most of their time in school, you know, being graded on their individual performance, right? You know, they’re—not based on their work together as a team. They’re being judged on their ability to answer questions that, you know, already have answers. And then they come to Snap and we’re handling them our big problems that have no answers.

And I think—to your point, which was very well made—I think there is not necessarily enough of a focus on all the socioemotional skills that allow people to work together on a team. And I think, you know, now, just given the rate of technological progress, it’s almost impossible to do anything interesting unless you’re on a team. Because, you know, the work today is so multidisciplinary. It requires people with so much expertise in so many different areas that unless you’re attuned to working together as a team, it’s going to be incredibly difficult to succeed today.

So I totally agree that we need to think about pivoting education to focus on these things, like, you know, team-based learning, like evaluating people based on the work they do as a team, like helping them tackle questions that don’t have answers, and using that sort of creativity. And, of course, things like socioemotional learning. So I think all of those things will be real assets in this new era. But I also think we should bring AI into the classroom because it can provide personalization. I think one of the biggest, you know, challenges we’ve had in our education system is that everybody learns the same things at the same pace. And I just think that doesn’t reflect at all people’s individual capabilities or passions. And I think, ultimately, we’re going to see, you know, a lot more human flourishing by bringing these tools into the classroom, rather than pushing them away. And then by empowering people with the skills they actually need to be successful in the, what, 50 percent of jobs that haven’t even been invented yet that folks will have in the future.

NICHOLS: Would you go back to the audience? OK, right in the front here, in the—I hesitate to describe anything. The brightly colored sweater. The madrigal sweater.

Q: Hi. Hi, how are you? My name is Carmem Domingues. And I work in AI and I study the ethical and societal implications of AI.

So I wanted to ask you about the glasses, or the Spectacles. What are your thoughts on the privacy and digital surveillance implications of it? Not just for the person choosing to use it, but those around them who are now susceptible to, you know, be recorded 24/7 without their explicit or implicit consent?

SPIEGEL: Yeah. I think this is a really thoughtful question. It’s something we learned a lot about in the early generations of our glasses that were just camera glasses. And we really focused on having very, you know, large outward-facing indicators if content was actually being recorded. So if you’re making a video, having a big light so people know that you’re being recorded. My guess would be that in the future generations of these devices a lot of that processing will happen on device, and that there will be a very large degree of anonymization that happens for any content involving other folks around you. Rather than a device, for example, you know, recording streaming video frames and saving those frames forever.

So I can’t speak to what other technology companies are working on or doing, but we see a lot of technical paths to solving that problem. You know, not only by showing people if they’re being recorded with an outward-facing indicator, but also really thinking about how—you know, if folks are, you know, collecting context, for example, in their environment—how that can be done in a privacy safe way, you know, including anonymization.

NICHOLS: And right here in the front, and then we’ll go right behind you. So you’re first, yes, in the sports coat.

Q: Hi. I’m Nick Kazvini-Gore. I work at a company called Applied Intuition.

I had a question about, sort of, the transition from the kind of the age of social media, and now a lot of engineers when they come out of college want to work for companies that are working towards the national mission, whether it’s defense technology or, you know, like, drones, looking at utility, power lines. So I’m curious if you view that as, like, a continuation from the work of, like, social media, and now into defense technology, and your thoughts generally on, like, AI and national security. And if you see your company having a role in that, or not. Thank you.

SPIEGEL: Yeah. So, first of all, I’m really excited about that trend. I think that’s a positive one. I think part of it is the national mission, but part of it also is that capital for consumer technology companies essentially dried up, right? And so there weren’t a lot of opportunities, I think, for folks who were more interested in the consumer tech paths. And there hasn’t been a lot of distribution for new consumer tech ideas. I think, you know, that distribution, those opportunities tend to be created in these big platform shifts, right? Like the emergence of, you know, the desktop, or the internet, the smartphone. I think, you know, AI may create some of those opportunities. I think glasses will as well. But I think that’s usually when you see that that growth in consumer tech companies.

So I think a lot of that, frankly, you know, it was sort of a combination of a recognition that we needed to invest more in American leadership and in defense tech, and other sort of hard technology areas, combined with a lack of capital, frankly, for new consumer technology ideas. So for folks who wanted to build something coming out of school, I think defense tech has been a great place that a lot of that talent’s been absorbed, and I think, you know, be a net positive overall.

NICHOLS: And then right behind you, who overlooked in the past.

Q: Thank you. Guillermo Christensen. I was the CIA fellow here a long time ago, and now practice law.

Another device that seems to be on the cusp of something interesting are these small, wearable devices that record everything you do all the time, and presumably bring your life into a digital realm. Wondering if that’s something you guys have looked, or your thoughts about something that captures everything all the time.

SPIEGEL: Yeah. I mean, you know, half of me thinks you already have that with your smartphone, right? Small device recording everything all time. And then, you know, the other part of me thinks that, you know, there’s limited utility today in those types of devices. A, because the interface is primarily an audio interface, which ultimately is just very low bandwidth for humans, right? We’re very visual. Our visual cortex is incredibly powerful. We consume information visually. So I think devices that only have audio input and output, or even a small screen, are just going to be limited in their utility overall, and ultimately, you know, generally do tasks that the phone is capable of.

I am not yet convinced that there’s a ton of utility in recording, you know, your life all day every day. I think there’s probably more utility, and I’ve seen a lot more companies in this space that are interesting, where they’re just recording your use of your computer all day, every day. That, I think, is quite interesting. And I think a lot in that space is going to, you know, create new opportunities to create proactive recommendations, or use your computer for you, with probably fewer privacy concerns. So that’s the space I’m watching more closely, I would say.

NICHOLS: Please, right here.

Q: Henry Gundlach.

My question was, has Snapchat benefited from ads shifting away from TikTok, given it some uncertainty over the past year? And if so, to what extent?

SPIEGEL: You know, I don’t think so, to be quite honest with you. You know, maybe to some degree here or there. I just think in general advertisers want to diversify. That’s a message that we constantly hear from them. And I think TikTok is part of that diversification plan. I think we’re part of it as well. And certainly some of the smaller folks, like a Pinterest, or Reddit, or something like that. But we just constantly hear from advertisers that they want to—they want to diversify. They can reach new audiences, drive new conversions, that sort of thing.

NICHOLS: So you don’t see it as a competition?

SPIEGEL: It’s certainly a competition for advertising dollars. But I think if you look at where the dollars are today in the ecosystem, right, they’re with Meta and Google, right, or Meta and YouTube, you could say. So I think for us, as we look at, like, where we can win more share, we’re less focused on winning share from TikTok, for example. We’re more interested in winning share from Meta, for example, just given their scale in the market.

NICHOLS: It’s just a bigger pie, yeah. Please, all your questions are better than mine. So right in the front here.

Q: Hi. Doug Ollivant.

My son’s an Otis grad. So I’m perhaps more aware than aware than most of what you did for the class of ’22, forgiving the student debt for that entire class, for the most part. Can you talk about what the future is for tech philanthropy? How do you think the founders of these countries (sic; companies) can best use what can only be described as obscene wealth to, you know, help America, its people, how it’s going to advance future legacies, and so on?

SPIEGEL: Yeah. What a fabulous question. You know, I definitely think there’s a legacy in the technology industry of, you know, substantial giving. With things like the Giving Pledge, for example. I think, you know, we’ll see a whole new wave of, you know, folks signing up for the Giving Pledge, hopefully, kind of as we get through some of the growth of these new AI companies, which should create quite an enormous amount of wealth. I think personally, for our family, we just feel an obligation to give back. We’ve been incredibly blessed with so many wild opportunities, probably most importantly being born here in the United States of America. And so we really tried, you know, where we can, to give back meaningfully. And just will—and will continue to do so as a family.

NICHOLS: Right in the front here.

Q: Thank you. Paula Stern.

And I don’t want this to be a gotcha question. I’m really interested in your insight. I’ve been working for a long time, since the National Science Foundation set up an organization which it gave the name of the National Center for Women in Information Technology. And that word, “women,” has meant that basically DOD’s dropped our funding, NSF, because of—whatever. I’m interested in your insight, both within your company, the role of women computer scientists, information technology engineers. What is their contribution? And does it make a difference whether there are women that are at an underrepresented rate, about 24 percent overall, in this arena? Just very interested in your observation. I know you’ve got four boys, et cetera, but think about the other half and tell me what you think.

SPIEGEL: Four boys and an amazing wife, so very fortunate to be married to an incredibly accomplished woman. I think making sure that we are finding the absolute best talent all across the country, and the world, is an imperative for our business. And we have found that, you know, the more that we look for folks, sometimes even in unexpected places, we’re able to uncover just extraordinary talent. So we invest a lot of time and money trying to find amazing folks from all sorts of different backgrounds, because all of those different perspectives are very valuable. I mean, that’s one of the key drivers of innovation, right, weeing things from different perspectives, having those sorts of conversations. We are fortunate to have some really, really strong female leaders at Snap, who I’ve learned a ton from personally.

So I do think that it’s incredibly important that we continue to find great talent all over the place. And I do think women have played such a pivotal role in computer science, in particular, since the very early days. So I think they will certainly continue to play an important role in the future. And, you know, it’s imperative for our business to make sure that we find the absolute best talent, you know, wherever those folks are.

NICHOLS: OK. I believe we have another virtual one. OK.

OPERATOR: We’ll take the next question from Tony Pan.

Q: Thank you, Evan.

So I’m an older millennial, so curious about the young. Given Snap’s dominance of the youth demographic, Snap’s chatbot, My AI, is probably the very first generative AI that millions of young people regularly talk with. So, just curious, what has Snap learned about how the next generation interacts with AI? And how is it different from, like, old people like me? (Laughter.)

SPIEGEL: You know, it’s a good question. It hasn’t been particularly surprising for us, although we’ve put a lot of guardrails on the way that that My AI works. We partner with folks like Open AI, like Google to power the chatbot. But then we add a lot of guardrails and protections around what it can do. So it’s generally used for general knowledge questions, which probably would be very familiar to you as a millennial. So unfortunately, probably not a lot of unique insight into the way that My AI is used. Although one of the things that’s been fun is that we made it really easy for you to bring it into conversations with your friends. So if you have a question or you want to fact check a group chat, you know, or you’re looking for a fun place to go out with your friends, easily bringing My AI into those conversations has been fun.

NICHOLS: What’s the most embarrassing question you’ve asked ChatGPT? (Laughter.)

SPIEGEL: I don’t know if it’s embarrassing, but I love that, like, because I have so many—you know, I’ve written so many things publicly and spoken publicly so many times, it knows my voice very, very well. Which really comes in handy for pretty much anything, you know, I do. And I think what I really love to do with ChatGPT is run things through simulated focus groups. So if I want to, like, learn something really quickly or get a gut check on something really quickly, it’s so fun to have it. You know, you can build a simulated focus group of really anybody you’d like, and, you know, get a gut check on something you’re thinking about, or a new product idea, or a way of communicating. And I think, you know, getting that feedback really quickly and early has been fun for me.

NICHOLS: In the back there with the glasses and the—yes. Yes.

Q: Hi. My name is Sherry Hakimi. And I work with pirth.org.

My question is about trust and safety. So right now you mentioned that anybody who’s on Snap, if they feel threatened or if there’s any sort of harm they can immediately get help from the trust and safety team. Not all platforms make that as easy. And more often than not when people are getting harassed, or cyberbullied, or threatened it’s happening across multiple platforms. So I’m curious, is there a way that the platforms might be willing to work together to build a universal reporting system that would make it easier? And if not, what would it take to get everybody on board?

SPIEGEL: Yeah, that’s a great—it’s actually something we’ve invested in a lot, to do cross-platform signal sharing. Obviously, it’s gotten a lot harder because the platforms, especially with the Apple privacy changes, have removed the ability to, you know, detect the unique ID. But we can still match other identifiers. So this is actually something that we do. We do with Meta and other companies. And it’s, I think, a really promising industry initiative. I do think companies take this very, very seriously. And, you know, fortunately, with a lot of the tools with AI and a lot of the investments that we’re all making as an industry, I think these platforms will continue to get a lot safer. So, yeah. You know, if you’re interested in more information, I’m not sure what we’ve shared publicly, but that is a big industry initiative at the moment.

NICHOLS: We’re going to do one final one. We’re going to give the prerogative to Mr. Spiegel if he has a question for his son. (Laughter.) And if not, I’ll take one more myself.

Q: I’ll pass on the question. They’ve all been great.

NICHOLS: Well, they have been great. I will do—and sorry for your loss. I believe you lost your home in the L.A. fires. And I know that your family home. So I’m sorry.

Q: This was the home where Snapchat was invented, actually. And I was no longer living it.

NICHOLS: Yeah. Devastating.

You’ve given us so much optimism here tonight. Give us the most optimistic case for how L.A. rebuilds and looks in the future.

SPIEGEL: The federal government gets so excited about the upcoming Olympics, and World Cup, Super Bowl, I think, in L.A., that they decide to make a huge CDBG-DR for community redevelopment. We’re able to actually disperse those funds very quickly to the people who need it, and to rebuild the necessary infrastructure. That L.A. gets really inspired by all the advancements in AI-based permit approvals and is able to process those permits very quickly. And so, in addition to the fifteen or so houses I think are being built right now, just in the alphabet streets in the Palisades, we see folks putting up houses everywhere. And they do it in such a safe and thoughtful way that they can actually get insurance at the end of the road, because the community is much more fire resistant. And then, through it all, L.A., especially both sides of LA—because I think these fires have connected to, you know, both the community of Altadena, of course, Palisades, Malibu—connect these communities across L.A. in a way that they haven’t been before.

NICHOLS: Well, thank you very much. I hope everyone found that conversation kind, creative, and smart. Evan Spiegel, thank you for joining us. (Applause.)

(END)

This is an uncorrected transcript.