Van Jones and Arun Gupta on the new AI civilization

Hey, welcome to Politico tech. Today's Thursday, May 9th. I'm Stephen Overle. You're going to hear two guests on the show today. The first is Van Jones, CNN host and contributor, bestselling author and a longtime activist who now runs the Dream Machine Innovation Lab, a nonprofit using technology to uplift neglected communities. And joining us is Arun Gupta. Arun spent 20 years at the venture firm Columbia Capital, and he now leads an organization called the Noble Reach Foundation. Its aim is to recruit tech people into public service. Now, the two of them first met at south by Southwest, the tech festival in Austin. And then I met them earlier this week for a live podcast taping at a Washington AI Expo hosted by the special Competitive Studies Project. What brought Van and Arun together is that they're both AI optimists. They say a lot of AI talk is too doom and gloom and that there are reasons for Washington and minority communities alike to, to embrace the latest tech wave coming from Silicon Valley. I had questions about that. Here's our conversation from the expo. So thank you both for joining us on a live taping of Politico Tech.

Love it.

Thank you. I'm going to dive into questions here. You all, I think, are working at a really interesting intersection between tech and government and a lot of the questions that frankly, are dogging our society about where AI is going. Van, I want to come to you first. You recently described the people developing and shaping artificial intelligence as, quote, co founders of a new human civilization. Why does this moment to you feel like a new beginning?

Well, I mean, let's just be practical about it. I have four kids, but my little kids are very little. I have a 14 week old little boy. He is going to grow up in a different human civilization than I did. His first crush is probably going to be an AI. It's not going to be a pop star. It's not going to be a rock star. It's not going to be a cartoon character. It's going to be an AI that can interact with him, might know him better than he knows himself, better than I know him. That's weird. He'll be able to use biotech tools to design my grandkids. Conceivably, he might be buried on Mars or the moon in 80 or 90 years. So that is. You think you're founding a company, you think you're founding a fund, you're co founders. We're co founders of a new human civilization. And the things that we put in place for fairness, for opportunity, for safety, for security are going to last and have impact for generations to come, and we should take it very seriously.

Arun, I think perhaps the most influential co founders of this new civilization will be the people actually building this technology, right? Determining how it's made, how it gets deployed. How does the approach to AI feel different from what you've seen in past waves of technology, or is it different from what you've seen?

Yeah, it's a great question. I mean, I, I think what's different about AI is that I think there's at least an understanding of the gravity of what's possible, both good and bad. Right? I think, look, when we, when Internet came around, we didn't really know how it would change life, we just knew it could. Now we have models to say, here's the really good stuff that happened and here's some really bad stuff that happened. When social media came around, it was all about connecting friends and people. Now we understand there's some really good stuff that can happen, but there's also some really bad stuff that can happen. And I think the lesson hopefully that we're learning with AI, because arguably, I agree with Van, it's more transformative than either of the two things I just described, is that we at least start to think about the art of the possible, of what are the possible good things that can happen, right? And we need to focus on that while putting guardrails. In our country, we're usually reactive to putting guardrails in, like, until something really bad happens, then we go like, oh, we need to put a guardrail in, and then we over regulate or the pendulum swings too far. But how we can be maybe proactive about putting some of those guardrails in early. And so I think what AI has the benefit of is having transformative technologies that have come through that actually can at least inform the discussion in a way that we didn't have before.

That's when I got into tech coverage. I think of it now as like the gee whiz days of tech, where everything was this cool app, look what this new Twitter thing can do. And now it does feel like a more nuanced conversation, right, that you have to recognize these things. These technologies can have negative changes too, and you have to think about that on the front end. Van, with that in mind, how confident are you in the people developing artificial intelligence to go about that in a different way, in a more responsible way than what we've seen in the past?

It's too early to tell. I mean, I think people forget a lot of the people who are working on These large language models and stuff like that. They're shocked themselves by how powerful these things are. They're shocked themselves by what's emerging. And so they're now going to have a lot of power. Power reveals. It's one thing you learn here in Washington D.C. you think, you know, somebody, get them elected into office, give them a lot of power, and either the good or the bad will come out. So we don't know and I don't think that we should wait to find out. In addition to whatever government regulations could be put in place, I think it's important for communities to lean forward into AI. A big concern I have is that people are being afraid to even try, which we're kind of killing hope before people have even tried. I think the communities I care about need disruption. If you're living in a low income, overlooked, underappreciated neighborhood, that school system needs to be disrupted, that healthcare system needs to be disrupted, the opportunity envelope needs to be disrupted. I'm not afraid of disruption. I'm afraid that the people who are doing the disruption right now have all the power to disrupt, but no purpose. They're just disrupting for its own sake. They're disrupting based on greed and speed. They want to beat the other AI company, they want to make the most money. That's not a good enough set of values, intentions and motivators when you're going to literally be laying the groundwork for a new human civilization. So we need each other. You need the people who have a purpose for disruption but no power, and the people who've got power to disrupt but no purpose. You put them together, the best of AI, the best of community. You might start solving problems at the grassroots level that were never solved before. You put real data, real power, real computational ability in the hands of some of these grandmamas and these coaches and these reverends have been on the front lines for a long time, they might start solving problems.

Just a follow up for you on that is who's responsible, you think, for making that happen. Because there are communities that have been left behind by past tech waves or have been burned and have reason to be skeptical, I think, of artificial intelligence.

It's a three sided problem. The companies have to be more than willing to engage with communities that can't write a big check. The corporations can write a big check, militaries can write a big check. So the tendency is for these companies to try to serve that. They are pulling in trillions of dollars. They should be dedicating a certain amount to make sure that they are responsive to people who can't write that check. That's the responsibility of the companies. Communities can't be so damned afraid of everything that we assume the worst before we even try. Sometimes we psych ourselves out. It could be the best thing ever happened to poor people is AI. It could be that young people are going to be able to educate themselves better. Businesses in those communities can cut their costs. Less accounting costs, less marketing costs. It could be great or it could be triple. We don't know. So communities got to step forward and then government's got to make sure that there are rules that make sense and they're applied fairly. But we can't be so afraid of AI that we throttle innovation, let China beat us and leave out those communities. So it's a three way conversation and that's a part of, you know, us@dream machine.org we're trying to make sure that three way conversations happens the right way.

And just on that point, if I could just add, to reinforce it, you know, chatting with a senior official at OpenAI that went and did a 30 country tour and you know, I asked her, you know, what did you take away from that? You know, and she's like, you know, I felt like there was an inverse correlation of just GDP and how opportunistic or how hopeful people were about AI, meaning the poorer the country, the more they were like, hey, let's talk to us about what we can do with AI. And the richer the country was like, whoa, how is it going to harm us?

Right?

And I think this is an important conversation, right? Because AI can be a real leveler if we let it be. And so I think the role that Dan's talking about on this three sided coin is that we need the folks developing AI to be enabling these communities to experiment with it. Because the truth is, and as was said earlier, no one really knows. You ask folks at OpenAI, they can't tell you why the models do what they do. Completely this idea that, hey, tell us what algorithms you have, that feels good, but no one really knows why the algorithm still act a certain way. Right. The only way we're going to know is experimentation. And we need to encourage experimentation in all communities.

I want to talk about the role of government in this dynamic, right? Where is the government's role to facilitate those experiments? Is it to put rules in place so that people can trust the technology and be more optimistic about it? Arun, what do you think?

Yeah, look, I think government had multiple roles, roles here, you know, look There's a policy, there's a technology, you know, at the core of it, there's talent, right? You can't be doing technology and policy if you don't have the right talent in. And so, you know, in the discussion we were having earlier, you know, I think an important part of government is getting the right AI talent inside of government so that you can start having these informed discussions right now. As far as the role that they would have, I think one, government has a role of experimenting with AI on itself. Like how can we be more efficient, how can we do things and how do we provide better services to our communities? Leveraging AI, that's 1, 2, how do we think about new things that haven't been created that could really help our communities in a different way? So value added services, that's the optimistic, hopeful side. But then their role is also like let's think about the negative things from the conversations that we talked about earlier. What have we learned from social media? What have we learned from previous tech waves to start putting some guardrails in place? And to me, those are the roles that government plays in this with AI.

And just a quick follow up for you because the talent question is a big one I know you're grappling with what is Washington's best sales pitch to get get industry AI experts to come to Washington?

I think the best sales pitch for Washington and it really depends if you're talking early career or you're talking mid career. But to early career folks, spend a year or two with us, understand the most difficult problems that we have, and then decide what you want to go do and branded as like you got selected in a special course to come in and serve your country and then with that you can kind of go create whatever you want on mid career folks. You know, we need to be having them know they have the opportunity. Storytelling is important, right? This isn't about just come into government, it's about what's the mission that you're able to provide. Right? So if you're talking to people about coming in and changing the course of climate at doe, or national security at dod, in the CIA, or changing the course of how we manage health, those are really inspirational kind of motivations. And I think that's what we need to be selling to get mid career folks to kind of come in.

Yeah, go ahead.

I see it slightly differently because you're a better person than I am. I think we should tell people if you want to get rich in AI, come work in government first.

No, I agree with that actually come.

Work in government first because you're going to see a message, you're going to see 100 different opportunities to optimize, to streamline, to make things work better. And when you leave, you'll be a subject matter expert in that. You can set up a company that come and fix what you just saw. So I think if you look at how AI is going to develop, one of the biggest customers is going to be governments. They're going to have, except for big, big corporations and very small consumers, the big chunk of capital to be deployed is government in contracts, not just government in terms of grants and subsidies in contracts. And so I think that what Arun is doing is actually he's going to wind up making a bunch of billionaires who've got big hearts. That's what happens. You get in government, your heart gets bigger. You're also, your head gets more callous from banking against like impossible systems and you want to come out and fix them.

Well, let me be the cynical journalist here and I'll take this question to you, Van. You know, if I am graduating with my PhD in AI development or whatever the degree is, I'm not even smart enough to know and I can make $60,000 or $70,000 working at the Department of Commerce or I can make 350, $400,000 working at OpenAI or Google. Why would I pick commerce?

I think the reason you pick commerce is long term. You're going to run past the people who went early, did the obvious. The thing about this game, it rewards people who actually add value. I'm not saying that Right now the PhDs in artificial intelligence are a commodity. But when you have a mix of experiences, you're more valuable, it's harder to replace you. And so I think it's a hedge against a changing world. You can guarantee they're going to be governments here. If you understand government. I have a job right now on television. I was never trained to be on television. I have a lisp, I'm hearing impaired. But I work for government for a very short period of time. And I'm golden for TV because most people don't work in the White House. And so I'm going to, trust me, anybody who takes his bet, who does the 18 months, who does the 24 months and comes back, they're going to be golden.

Yeah, I love this question because first you got to think of your career in a 30 year.

That part, right?

This isn't a two year race. Right. This is a 30 year race. Second, you're going to have multiple careers in this generation. You're not having one career, so you're going to have four or five careers. So. So what we're saying is have one of them be public service. Right. But not for public service sake. To go learn. Right. To go learn where the problems are, where the data is and who the people are and what the language is that you need. Right. Ultimately, where there's a short, there'll be plenty of folks going to AI companies to do AI stuff. The people that are going to get rich are the ones that are going to know how to apply AI to solve problems.

Van, you are working on these issues through the Dream Machine, right? Which is your organization. Jeff Bezos very famously gave you $100 million check. I imagine that was a surreal moment for you.

It happens all the time, right? All the time, yeah.

How are you spending that money to address this problem?

Well, we've allocated a good chunk of the capital on the traditional stuff that people expect me to spend money on. Criminal justice, climate, youth violence, et cetera. Very proud of those bets. We're going to see how they turn out. But we've actually reserved a pretty big chunk of capital for this. And we have launched a campaign working with Will. I am the campaign. We have any, like, Marvel superhero people, Marvel superheroes, nerds, Any nerds. Okay. Just. Okay. All of us. Good. So the campaign is called Make Wakanda Real. Make Wakanda Real. Wakanda being the. The African nation that's technologically advanced in the Marvel superhero universe. And the idea for this campaign is to get people excited about AI My concern is that most of the stories about AI are scary. Scaring the bejeebers out of people. It's all the Terminator and robocop, and we're all gonna die. And people in my communities that I care about are already depressed enough. We have enough bad news. This could be hopeful, this could be positive. But you've got to put the tools in the hands of the people who have the problems, who want to solve it. I'm predicting that you're going to have young people doing extraordinary stuff right off the bat. On June 15, we're going to do an AI opportunity summit in Atlanta. And what we're going to do is we're going to do. We're not going to be doing speeches like this. Maybe one or two workshops, hands on. Have the coach, have the student, have the teacher, have the grandma, have the owner of the barbershop, the nail salon right there, figuring out how to use These tools in real time to save them money, to create that resume, to create that business plan. Let's use this technology for good. Let's find the companies that want to allocate. These companies got to allocate some money to help. You can't pull in a trillion dollars and use it all to help militaries and corporations. There are other people who have needs. Some of this compute has to be set to one side. What about the black colleges? What about the tribal colleges? What about the Hispanic serving institutions? Institutions, put them at the front of the line. What's it going to cost you? A fraction of a fraction is billions. For these communities, this is the biggest opportunity that we have for something positive to happen and we're going to make Wakanda real.

I want to stick with the money topic because money is power. A lot of time, let's be honest. And venture capital has a lot of money, right? VCs pick winners and losers in tech all the time. And so if you want to invest and grow more black owned, Brown owned AI startups, VCs are a key part of that. If you want to invest in safer, responsible AI so people trust it, VCs are a key part of that. What does the venture capital community in your view need to maybe do differently than what it's done in the past? Now that we are at sort of the dawn of artificial intelligence and we, you know, as you're saying, you want to see all of this upside.

Yeah. So I think, you know, as we look at it from VCs, we're always going to be looking to play for the big wins. Right. Having said that, there's always in those big wins is encouraging those companies to be experimenting in these communities. Because these communities, when you aggregate them together are really material. You know, I've got a friend out in the audience whose board I was on, like take a simple idea as providing managed services to the government. You could do that in D.C. or you can do that in rural southwest Virginia. You do it in rural southwest Virginia and you create economic development and opportunities. You upskill people. So even how we encourage people to build their businesses matters. His company went on then to get acquired by a large company because they know they can't go do that. Company's 1901 got acquired and I think that's the experimentation that what Van's saying that we need to encourage and there are models out there. The problem is we don't tell these stories. We don't tell these stories in a meaningful way because the beauty of AI is the amount of capital that is required to run the experiment has drastically come down.

Yeah.

We can run 100x more experiments than we could have before.

Yeah. And the thing is, you can replace capital with code. That's what you want to think about. So you can replace capital with code. Stuff that was very expensive before is very cheap now, but if you combine it, if you actually make some capital available and make some compute available, we have no idea the amount of good that can be done. I'm glad that you're talking to us because I think we're hearing so much about algorithmic bias. We're hearing so much about it's going to kill everybody. It's going to get out of control. Listen, there are people working on that very, very hard right now. Most Hill staffers right now are aware of that. People are trying to figure out how to regulate it, but we're not figuring out how to unleash the upside. I'm a civil rights guy. I like equal protection from bad stuff. I don't want algorithmic bias. I don't want racist robots. Okay. I don't want that. But Dr. King didn't stop with equal protection from bad stuff. He also talked about equal opportunity to good stuff. I want equal opportunity, equal access, equal enthusiasm to the good part, to the upside. And we got it. We have. It's got to be better balanced.

I knew we're about out of time. So I want to end with one quick question. I'm going to put it to each of the Arun and then to Van, which is to reach this sort of future that you're talking about where we have AI upside, but minimize the bias and all the harms that we know. This technology can also perpetuate. What's the one thing Washington should be doing right now to get to that end? Goal New civilization you're talking about.

All right, I've thought about this a lot. Every time DC has a technology hearing, they bring a bunch of tech CEOs up here and just beat the hell out of them. I think they probably gotten the point that Washington D.C. is paying attention to the downside. Why not bring some of the people up here who are doing positive stuff with AI? There's something called Path AI. It was created by a young African American guy. And it's basically he recognized that people who want to get jobs and want to have their careers take off. If you've got a social network, you can do really well. But if you're just a kid from the hood, you don't have that coach, you don't have that mentor, you don't have that network. So this guy named Chandler, African American guy, created a company called Path AI. It's basically a career counselor in your pocket. And it gives people, not only does it give them coaching, it helps them negotiate their salaries. This is like right now happening. There's a guy who went to jail, went to prison, made mistakes, he came back home, he wanted to turn his life around. This guy goes to 50 job interviews and gets turned down for all of them because he doesn't know where to go. Who's going to hire someone like me? So he helped build a company called Untapped Solutions. And it's basically LinkedIn for people coming home from prison. Those are the people who should be brought up here and given the opportunity to testify about what they're doing positively. You've got to put a spotlight on the good instead of just putting a spotlight on the bad.

Yeah, I mean, look, I think this is why, you know, Van and I, you know, when we met at south by Southwest, you know, looked to collaborate, is that we're trying to lead with optimism. Right. I'm a VC by training, and I was teaching at a university after that. And the commonality, common thread, is optimism. Entrepreneurs are optimistic intrinsically, because if you're not, you're never going to get out of bed. And if you're 22 on a college campus, you better be optimistic because that's pretty damn good. Right? And I think we need a government that's promoting that optimism around what this can do. I worry that the only time we hear government and AI is about berating tech players around what they're doing. And the problem with that is we're still figuring this out. Having said that, we need guardrails, the tech players will say that we need guardrails. So let's start putting those in place from lessons learned in social media. But encourage experimentation. If there's one thing I would say government should be trying to do is encourage experimentation of AI with communities. Because small experiments, we don't know what we don't know yet. Right? So the more experiments we could be running as a country in communities, the better off we'll be.

Well, we are out of time, though. I could talk about this all day with you, too. Thanks for being here for a live taping of Politico Tech.

Thank you for having us.

Thank you all for joining us. That's all for today's Politico Tech. For more tech news, subscribe to our newsletters. Digital future daily and morning tech music in today's episode comes from the mysterious Brakemaster cylinder. Our managing producer is Annie Reiss. Our producer is Afra Abdullah. And our editors are Steve Heuser, Daniela Cheslo and Louisa Savage. I'm Stephen Overle. See you back here tomorrow.