AI vs. Democracy

[00:00:00] When I worked for President Obama, we said, let's go ask the American public, what should be our policies? What should we focus on in the first hundred days of the administration? 125,000 comments we got from 84,000 people. You know what we did with that? Zero. Because we didn't have the ability to make sense of any of that. Two of the hottest issues of our time are AI and the deterioration of democracy in America

[00:00:25] and around the globe. But have you ever wondered how these two issues are linked? Could it be that AI has the potential to save democracy, to renew it in powerful ways, even though it also has the potential to destroy it? In this powerful episode, we'll be talking to one of the leading thinkers around the globe on the role that AI can play in the renewal and in the destruction of democracy. You don't want to miss this episode. Let's get started. Hi, I'm Dr. John Izzo and I want to welcome you to this episode of The Way

[00:00:55] Forward Regenerative Conversations podcast, where we tackle the big issues of our time and ask how can humanity and planet have a positive way forward. We welcome you to this episode. You know, over the last few months of this podcast, we've been doing a series of conversations about democracy, polarization, and really how we find our way forward, especially in the United States, but really around the globe. And the interesting thing is if there's one thing Americans seem to agree

[00:01:23] on these days, which aren't many things, at least according to the polls, across age, geography, and political identity, we do agree on one thing, that our democracy feels broken and that the model we've relied on for generations is no longer delivering what people need. Some interesting statistics that might really put an underline. A 2024 Chicago Council survey found that 67% of Americans

[00:01:48] feel that weakening democracy poses a critical threat to the United States. Even more interestingly, science news found that nearly three quarters of Americans felt that American democracy used to be a good example for other countries to follow, but is no longer a good example. And a PBS AAP poll showed that only about one in 10 Americans say that our democracy is working well.

[00:02:16] As many of you know, I'm a senior fellow at the Stimson Center in Washington, D.C., where my work focuses on reimagining democracy, how we renew the civic fabric, rebuild trust, and create institutions that actually work again. And I'm just back from the big TED conference where I got to attend a few weeks ago, where the dominant theme was AI. And depending on who you listen to, AI was either about to destroy

[00:02:41] the world or save it. But what struck me is how little of that conversation focused on democracy itself, when I believe AI may help us govern better, listen better, and solve problems more effectively. That's why I'm especially excited about the guests that we have today. Beth Novak is one of the world's leading thinkers on the intersection of technology, governance, and civic innovation. She served in the Obama White House, founded the GovLab at

[00:03:08] NYU. And her new book, Reboot AI and the Race to Save Democracy, is one of the most important contributions, I think, to the conversation around AI and democracy. And today we're going to explore those issues in detail. And by the way, if you're not following me on Instagram or TikTok, Dr. John is stepping up. I am regularly, almost daily, commenting on issues of democracy and polarization. So Beth, welcome into the WayForward studio. We're glad to have you today.

[00:03:37] Thank you. I'm delighted to be here, and especially with you, and especially to have a, what we hope will be a hopeful conversation after that rather depressing intro. Yeah, exactly. And although, you know, it's depressing, but as someone who's spent his whole life on systems change, I always say it's kind of like an alcoholic or an addict. They say they don't really change until they hit bottom. So while those statistics are depressing, I think the good news is

[00:04:04] there's a hunger for change that's emerging. And of course, hunger sometimes leads to eating, and we'll hope in this case that it does, meaning eating some positive change. So Beth, before we get into your book and some of your great ideas, I want to ask a more personal question. What got you into this conversation? Is there a story or a moment in your life or just, you know, an experience that

[00:04:29] really catapulted you into wanting to be involved in this particular issue around AI and democracy? Well, maybe let me pick up on the eating hook, and I'll start with apple strudel. I had the wonderful experience, both during university and after, of spending a good deal of time in the coffeehouses of Vienna, eating apple strudel and a whole lot of other cake, I might add, but also studying what

[00:04:57] was it that happened in the 1920s in Europe, in Germany and Austria in particular, where you had these fledgling, very hopeful democracies that everybody was excited about after World War I, that in many cases devolved, as we know, into fascism, and then ultimately World War II. So this question of what made some political systems and democracies thrive and flourish, as was the case here and in other places like Germany and Austria, we saw institutions that were brittle and they broke

[00:05:26] and succumbed to fascism. So that's the earliest, you know, sort of origin of this in some way. But let me fast forward. I'm a child really of the internet, of the commercial worldwide web, the first generation that really sort of grew up with the web. And so the fascination that I think seized a lot of us and continues to, where we asked the question later, could these platforms, and you may remember this, people may remember back to when the web first started, we were really excited

[00:05:55] about this being the new civic town hall, this being this great place of free speech, and it would revolutionize our democracy. So I got involved very, very early on in building what was actually the very, probably the very, very first platform for community participation and democratic deliberation. We called it un-chat as opposed to chat. It was meant to be a place of civic dialogue and conversation. And of course, all of that hope has not all been realized. Some people refer to today's internet

[00:06:25] and social media as democracy's dumpster fire. We see the debates around people wanting to ban cell phones in schools, the lamentation about what social media and technology is doing to the mental health of our children. So I think it is promised unrealized. But that's my origin story. It goes back to the eating and the strudel, and really just this fascination with what makes some institutions work, and what makes other ones break.

[00:06:52] Yeah, I want to come back to that in a moment, because having studied that, I'm curious what couple of insights stand out to you. So while you're thinking about that, you were talking about the internet. You're right, when it first came, you know, it's going to bring the world together. We're going to be able to connect with people all over the world. And of course, there have been moments like that, the Arab Spring and, you know, things like that. But for the most part, agree, you know, and the research would agree, failed project in terms of bringing us together. And one of the exciting things

[00:07:20] to me about AI is we get a chance to do it again. And the interesting thing is, this is kind of the internet on steroids, isn't it? And what I mean by that is, in both directions, it could do even more damage to democracy than the internet did. And it also could do a lot more good than the internet had the potential to do. But let's hold off for that just a moment. Tell me a little bit, in sitting

[00:07:46] there eating the strudel, what's a couple of things that stood out to you about the lessons that that maybe are especially relevant at this moment in American democracy and in democracies generally. Okay, I'll leave to one side, but you're the best coffee houses. That's another podcast. But what emerged really from that conversation, and from looking at the rising tide of political violence

[00:08:10] after World War One, was that the threats to democracy came not just from and you saw a very sort of, obviously, ultimately, it was a sort of right wing fascist movements. But you had this increasing rigidity from both the right and from the left. You had political parties who were vying for dominance, vying for control. Interestingly, each starting their own newspapers, each starting their own

[00:08:36] social organizations, each starting then their own militias, and competing for control over social and political life. So I'm reminded of how today, you know, most Americans, if you ask them will gladly marry somebody of another religion, but they don't want to marry somebody of another political party. We see rising tide of political violence on the left and on the right with it isn't a week that goes by now where we don't have some kind of an assassination attempt or an attack. If it's not against the

[00:09:04] president, it's against a judge or a local political official, 58% more tax on members of Congress. I think we're aware of this heightened climate of violence. And some of this, at least in the 1920s, was the result of this increasing sort of painting and vilification of the other side as the enemy. You got to join my political party, because the other guy is the bad guy. And we see a lot of

[00:09:32] this being repeated now. And it's why you have jurisdictions where you have independent candidates who are succeeding because people are tired of that sort of they're tired of both political parties. And they don't feel either one really speaks to them or for them. They don't feel represented. But also the manipulation of political discourse to really paint the other guy as the enemy.

[00:09:56] We saw a lot of that really leading to just sort of incredibly brittle institutions that combined with economic shock and hyperinflation, but this rising tide of partisanship that really led to essentially institutions that broke instead of bending. There's a lot more to be said about, you know, changing role of women and sexuality and other things that we could get into if this were a podcast

[00:10:20] about history. But long story short, I think the recognition of some of the same hallmarks that we're seeing today about rigidity and hatred and just sort of armed camps entrenched against one another, and the realization that we have to get away from that, that we have to get towards places where we are actually more collaborative, that we are working together, that we are thinking about how do we use these tools to bring us together, as you said, instead of dividing us, as politics seems

[00:10:50] to do. It's breakfast time here on the West Coast. You got me thinking about strudel now. I was in Vienna this past summer having strudel, so I get it. So let's come back to, you know, you said something I think is quite profound and true, which is that many people today would be okay marrying someone of another religion, which in the previous generation was kind of anathema for many people, or, you know, families would go crazy. But now to marry someone outside of your political persuasion,

[00:11:19] one could almost argue that in the United States in particular, that politics has become a kind of religion now with many of the same characteristics of religion, which is this, you know, belief that we have the truth and they don't, this belief that the others somehow is, you know, a heathen or an apostate. And I love your point that rigidity is one of the qualities that then became the soil for

[00:11:48] this kind of transformation. But let's get into this AI piece for a minute here. Give us a headline about your perspective on AI's potential. You know, in your book, you say something that strikes, you say, AI won't save democracy, people will. But what in your mind is the headline around the potential that you see AI having to change some of the dynamics in our democracies?

[00:12:14] So the first headline is that we have the wrong headlines. You know, headlines, newspapers like, as they say, it bleeds, it leads. So it is very popular to write now about one of two things, either on the one hand, the robot apocalypse, AI will eat all the jobs, AI will destroy humanity, as we know it, tech is evil, fill in the blank, terrible bad news. Or on the other hand,

[00:12:39] we have these headlines about the AI industry, AI jobs, anthropic is going to go public, open AI will be right behind it, we're going to have incredible wealth that's created, we're going to see American dominance, vis-a-vis China or vis-a-vis Europe. So you have these two poles, these two extremes, which is not atypical for any new phenomenon, but also because, you know, people want to sell newspapers. Again, not dissimilar to what was happening in the 1920s,

[00:13:06] where people wanted to do the same thing. And you saw proliferation of newspapers, all of which sort of emerged in order to sell that viewpoint of one side or the other. So the headline is really somewhere in this messy middle. And the same is true about democracy, and it's true about jobs. It is that these are powerful new information processing tools, and we can use them in order to solve the challenges of our democracy, which fundamentally

[00:13:33] come down to how do we solve problems in the world, in a world in which we don't have a scarcity of information as we used to have. We have an abundance of information. We have tons of content, tons of know-how. We're all connected all the time. And AI can really help us get through this sort of overwhelming mass of information to find our way through, to frankly address some of our hardest

[00:13:59] problems, if only we use it that way. So the short one line is, we need a moonshot for democracy. We need a focus on how we're using AI to address democratic challenges, not just talking about how do we use AI to solve climate change or improve wellness, but really how are we using it to fix our democracy itself. And it seems to me if we're going to talk about the fix, if you will, or the potential way AI could be

[00:14:28] helpful in that fix. It's good to say, well, what are the issues? And one of them you've already identified, which is with this rigidity, the inability to find common ground and to, you know, polarization. Another issue that you talk about in your book that we talk a lot about here on the podcast is the current relationship between the governed and the governing. And, you know, you make the point of the book that we often make, which is that, you know, democracy in many ways has been

[00:14:56] reduced now to voting. You know, we kind of vote, you know, every couple of years, you know, we have two choices in America, basically, for the most part, we vote. And then, you know, a few years later, we get to vote again, you know, rather than a really more organic ongoing relationship between the governed and the governing. And it seems to me AI has a potential for both of those things around polarization and also around changing the relationship between the governed and the

[00:15:24] governing. So let's start with each of those. What's the potential for AI around the common ground around? And maybe if you have some examples of where it's already being used, that'd be great. Absolutely. Examples are my favorite part. And that I think was one of the great fun things about writing this book, just to, you know, we're talking a lot about a lot of very heady issues here around democracy, which can sound very abstract to people. And we should probably get into exactly

[00:15:50] what democracy means. But the book and the joy in writing, it was writing about very real people doing boots on the ground work to improve the way that government listens, the way that we talk with one another. So let's get into some of those specifics and come back to your kind of common ground question. And this fact that most of us today really have no voice in how government functions. We may, you know, shout into the wind on social media, but ultimately other than that,

[00:16:19] once a year appearance of the ballot box, there isn't much happening. And there's a reason for that. So even well-intentioned people who want to say, oh, I want to go out and listen to citizens. I'm going to go out and have a town hall, right? How often does that happen? Once a year, twice a year, a bunch of people shouting at one another. When I worked for President Obama, to give you a concrete example, we went out and for the first time we said, let's go ask the American public,

[00:16:45] what should be our policies? What should we focus on in the first hundred days of the administration? So not surprisingly, he was the first sort of really organic internet president in some ways. 125,000 comments we got from 84,000 people, if I remember the number correctly. You know what we did with that? You know how much listening we did? I don't think it takes a rocket scientist to answer

[00:17:10] this rhetorical question, which is zero, because we didn't have the ability to make sense of any of that. And the average kind of government engagement or citizen consultation, you know, the UK government has actually done the math on this. They go out and they do do regular by law. They have to go out and ask the public about quite a number of policy things. But it can take months and months and literally millions of dollars for them to actually, you know, we're talking like people's full-time

[00:17:36] job for months to actually do that analysis. Okay, fast forward, what's happening now? Now in the UK government, they built something called the consultation analyzer. And what used to take months and millions now takes a matter of days and even hours. In the city of Hamburg, they have built a toolkit, not just to allow the public to comment on every time they make an infrastructure decision or

[00:18:02] environmental decision, like where should we put this new train station, or we're going to dismantle this old train station and turn it into like a park and restaurant, sort of a food hall. What should we do with it? What does the community want? Now, because of the tools that they've created on the back end to allow them to summarize and synthesize what the public is telling them. Now, every single decision that they're making is done with public input. Closer to home in California, the Engage California project

[00:18:29] started as a one-off pilot. Let's go ask people in California what we should do after the wildfires to do a better job with wildfire response. Because they had AI, which allowed them to go out and make sense of that information quickly, they were able to turn the ideas that the public suggested into policy and now go out and do it again and do it again and do it again. So what I'm really excited and hopeful

[00:18:55] about is that we're moving from a world in which we had no conversation between government and governed to a world in the web where we did it, you know, once in a while as a pilot or as sort of a showcase thing. But, you know, really, it was just too time consuming and too expensive to now a world in which I'm hoping over the next few years we are going to really build the regular infrastructure for making a more

[00:19:20] conversational government where we're actually having that discussion with people on a regular basis. I just got back from Capitol Hill. I was brought in by members of both parties, Republican chair, Democratic minority, genuinely asking the question, and I think they really meant it this time, we would like to engage the public on Capitol Hill, but we literally have no idea how to do it. It wasn't a lack of goodwill. It wasn't

[00:19:46] that people don't want to, they just don't know how. How could we use AI to help us was the question and where I think we now can answer that in the next few years and hopefully change what that relationship looks like. And I think that's hopeful and exciting. Yeah. And by the way, you should get the book if you're interested in this subject because there are many great examples in there. You know, one of my favorite examples that we've talked about in the podcast is, yeah, hold that book up. Go for it. I just, I just got my own copy for the first time. I got it literally this week for the first time. So I've got

[00:20:15] double duty. There it is. As one does. You got yours before mine. I did. Anyway, as an author, I did. It feels good. Somebody else got it before me. That's good. You know, in Taiwan, where they've been doing a series of initiatives using AI, you know, began with Uber drivers and taxi drivers who were kind of at war with each other. And we don't know what kind of policies to do, but through these AI tools that a lot of people don't realize one of the things can happen in

[00:20:42] these AI crowdsourcing tools. And there are lots out there now is that unlike traditional surveys, which is just one time, you just kind of go, okay, yes, no, you can do series of statements of agreement and disagreement and actually build in a way, a conversation that reaches a kind of commonality. So in Taiwan, what began with the Uber drivers and taxi drivers, they found all this common ground agreement that they would never have found

[00:21:08] without AI. So I think, you know, it has so much potential to, it's not quite like a conversation in a room, but maybe it's better in many cases. We're getting, we're getting closer to that. And even in places where we're having the conversation in the room, Mexico's Supreme Court decided to go out, as we know, back in the seventies, there was, there have been actually repeated instances of political violence in Mexico, the disappearance

[00:21:34] of university students, arguably by their own government. The parents never heard from them again. There's lots of concerns about cartel violence, as we know. Mexico's Supreme Court said, we want to go out and start to have conversations with our citizenry about our judicial system, about criminal justice in this country, with victims of crime, with the families of people who've been jailed in crimes, with prosecutors, with defense attorneys, and we're going to do it

[00:21:58] in the room. So they had 4,000 people that they brought into conversation around round tables, face to face, you know, with coming back to this, I don't know if there was strudel, but hopefully they fed people something equally delicious. Not strudel, I don't think. Not strudel. I'm thinking about even better things. But they used AI then by recording those conversations. Imagine you have 4,000 people sitting around a table. How many interns would it take you to then summarize and write a report

[00:22:26] about that? You're never going to do that a second time, but because AI, they were able to record it, you know, you put a phone in the center of the table, as we all know from our phones, somewhat annoyingly sometimes, it now summarizes your text messages, made it much faster and easier to make that transition from people talking to government actually listening. So in Bowling Green, Kentucky, they went out and they started a project called

[00:22:54] What Bowling Green Could Be, and they said to the citizens, you know, what do you want the town to be in 25 years? What's your vision for the community? And again, this ability to have new tools to manage that information overload, coming back to the original idea, too much information. We didn't really have a way through it. Now I can start to turn that into listening and learning organizations. And that's the big

[00:23:19] transition. So it can also be face to face. It can be online as they did in Bowling Green, face to face as they did in Mexico. But these great instances of people learning, how do I create that conversation? To your point, not so it's one and done. But so we're actually, you know, the kind of problems we're talking about, the hard problems in the world, you don't solve them in one conversation. First, I have to understand what's the problem. Then I have to go out and say, well, how might I solve this problem? And then even after

[00:23:48] identify solutions, I still have to say, well, how would I actually drive that through? How do we actually do that in Prague or in practice? So I think that opportunity to have repeated and ongoing conversations with people who are most knowledgeable about a problem, who are most committed in a community. That's what's really exciting and where I think AI, not AI, it's not AI that's doing anything.

[00:24:15] It's how we use AI and choose to use AI. And that's what I'm hoping this book can help us to just get smarter about is the things that we can do with AI. I hope you're enjoying this fascinating conversation with Beth Novak about the future of AI and democracy. We'll get back to it in just a moment. But I want to take a moment to thank you for being a part of our WayForward community. It means so much

[00:24:40] that you've joined us. Please subscribe on whatever platform you're on. It helps you because you'll hear about every episode and it helps us. Please comment so that you can be a part of the conversation. And we'd love to hear from you. Go to our website at wayforwardpodcast.com and let us know what you think and sign up there as well. And please, if you feel moved, support us. You'll see on our link tree

[00:25:05] how you can buy us a coffee, a virtual coffee, and support the good work that we're doing. So please, we'd love to hear from you. Please comment, please subscribe, and share this with others. That's how these great ideas get spread. Meantime, let's get back to this great conversation. You know, one of the other things you talk about in the book, which I think is so important, is so many times when we vote for someone, a party, an individual, whether it's locally or all the way

[00:25:34] up to nationally in countries, first, we may not know really what this person actually stands for. We certainly may not know what their actual record is. And then once they're in office, we often have no way to stay in touch with how they actually vote or do things. I remember Gordon Campbell, the former British Columbia premier, was asked once a question. I was there when he was asked this question that said, why do politicians break their promises? And I love it. He was running

[00:26:02] for the first time as premier, and I love his honest answer. He said, because you let us, because we make promises, then like three or four or five or whatever many years later, you vote us back in whether we kept our promises or not. But to let the citizens off the hook a little bit, the average citizen doesn't really know, you know, did Beth do what she said? What did she vote on exactly? And you were talking, for example, in Switzerland, in Germany, where there's an app now where people can, in essence,

[00:26:29] get a sense of their values, what's important to them, and then use that as a screen for candidates. And also some things that are tracking how people actually vote, not just the promises that they make. So talk a little bit about AI's potential and some examples of how it can really change our relationship with those running for office or running for re-election as well. Yeah. So even the electoral space and coming back to this sort of many different definitions of

[00:26:55] democracy, to your point earlier, you know, for most people, it's about voting once a year. When we think when you say the word democracy, most people think of elections. And we have had nothing but headlines either about actual subversion in our elections, actual attacks on our electoral officials, actual laws being passed that are making it harder for people to suppress their vote, make it harder for people to cast their ballot, or if not the reality, then the perception,

[00:27:24] the concern that there are risks and fears about our elections being in jeopardy. I'm trying to tell, not that all of that isn't true, and that there are a lot of risks and worries, absolutely. But there's a story we're not telling about how we can do better with our elections, hold our political officials more accountable, create opportunities for more people to vote. So let's talk about a couple of those examples. And let's start with the one that you raised.

[00:27:51] This one actually predates AI. In Europe, most countries in Europe, they have these systems, it's very simple. They come up each election with a set of here are the major topics and issues in this election. It could be education, it could be the traffic, it could be immigration. And they ask the same set of questions to every candidate. And then they use that to create a website. And so they ask voters the same questions. What's your feeling on how do you feel about immigration? How do you feel

[00:28:19] about the traffic? How do you feel about education? And then they essentially match you. They say, based on your answers, here are the politicians whose views most closely reflect your own. It's not really a very sophisticated technological thing, at least not in the sort of early versions of this kind of project. But this idea of simply saying we are going to make every, you know, politician doesn't

[00:28:47] have to respond, but these projects are really popular. And if you don't respond, voters aren't going to find you in the same way. So just by asking everybody the same question and putting it on a website, I give access for more people to that kind of question. Now let's fast forward to some of the exciting things happening with AI in the electoral space. This starts with the infrastructure of our

[00:29:08] elections. Groups like ERIC, which is an acronym that is the compact among 25 states, that's using AI to essentially clean up the voter rolls with an eye towards ensuring that people who have the right to vote will be on the voter rolls. We're not doing it in order to deprive you of a right to vote. States who are part of ERIC say, we want to make sure that Dr. John Izzo in this database, maybe you moved, you go between different jurisdictions, right? Maybe here you're called

[00:29:37] James and there you're called John, or here you're listed as, you know, J.Izzo, that you're not lost. And when you move, you know, you still have the right to vote. AI is just a better calculator. It's really good at finding patterns in data. And it's pretty good at doing this work of essentially spotting where we might have two people who sound the same. Is this person the same? Do they belong in the voter rolls? Then there's other sort of simpler stuff, like the fact that I can now easily

[00:30:02] generate text or create an image using generative AI means that as an under-resourced person who may not have the money to hire an expensive campaign consultant and a whole firm of people, I still could run for office and maybe compete on equal footing with someone who has access to that whole army of political professionals. Now, because my graphics are just as good as the other guy, I too can run for office, for example. I can do things like create training guides for the,

[00:30:31] you know, retirees who staff my polling place to ensure they get better instructions and better training on the day of, or give you a quick example of a project with my own students. The European example that I mentioned to you, we wanted to build, I wanted to work with my students and a state secretary of state to build something like that in the U S. But when we dug in and looked at that state's election website with that election division, we discovered, oh my goodness, that website

[00:31:00] was built in, I want to say 1990. Everyday content had been added to it since 1990. We couldn't build that kind of fancy A, what we thought we were going to build an AI chat bot to answer your election questions. But we realized the content was so out of date and was so old and so problematical. What we actually ended up doing was using AI to go

[00:31:26] through the whole website and find all the inconsistencies, all the problems and all the ways we could just simplify the language into something clean and clear and accessible to somebody without a high degree of education so that anybody could get better information, both about their right to vote, but also about how I run for office and hopefully democratize access to elections and render them what they're supposed to be free, fair and frequent.

[00:31:54] One of the things you say in the book, and because of piggyback on something you were just saying, is that AI in terms of democracy will only be as good as we are. I'm kind of paraphrasing you. And I think already just of two examples, and you'll see where I'm going with this question. Let's say gerrymandering. You talk in the book about how AI could be used to create more fair districts.

[00:32:18] And of course, wherever gerrymandering has been on the ballot in blue and red states in the United States, citizens say, we don't like gerrymandering. We'd like fair districts. So it turns out the citizens, whether in Ohio or California or Utah or Missouri, they want this to be true. But AI is also being used now to create districts and manipulate voters and find out exactly who's going to vote and how

[00:32:46] they're going to vote. So you see that both of those potentials there. And again, we could go on with examples. So I'm curious, what's your thought? This is kind of a deeper question, I guess. If AI and democracy are likely only as good as we are, what has to happen to change the dynamic that will allow that to happen? I was thinking about your visit last week in the Capitol, and we'd love to

[00:33:12] hear more about that. Because it seems to me, at least, that in the United States in particular, especially in the two-party system, that everything has now become about beating the other party. And ironically, Ezra Klein in his book about why we're polarized makes the point that the more we become a 50-50 country, it's actually more polarized than when one party held the Congress for decades, because you know you could win the next election, which means your total focus is winning the next

[00:33:41] election, not compromising because you know you could win and be back in power. So I guess what I'm asking is, do you have some thoughts on how can we as citizens push for this to change, given the dynamic that now exists in our country? Just an easy question. Yeah, there is so much to this question. So I want to state that, so there's so many sort of big picture elements to this, which have to do with

[00:34:11] the broader sort of how we got to this entrenched place of political polarization, that also have to do with the much bigger picture around ownership of these AI companies. And we have to recognize as much as it's easy for me to say, it's what how we use the AI, you know, and I use the analogy of nuclear energy and nuclear power, right? You can use nuclear to create bombs, or you can use nuclear

[00:34:36] to power our electrical grid, right? So it's, it's a bit of, I recognize a glib statement to say, it's how we use it for good or for ill. And that's the case for any tool and every tool. And partly, I say that in order to, number one, ensure that we lay the blame at the, in the right place, because anybody who says AI is causing the loss of jobs is disingenuous. It is CEOs, it is managers, it is

[00:35:01] people who are firing people. AI is not firing people, at least not yet. The reality of that happening, that's a ways off. It is people using AI as an excuse to save money at the expense of workers. So it's partly to ensure that we are just recognizing our own agency in this process. That said, there are a lot of big other political forces. If we have tools that are controlled by large companies over

[00:35:29] which we, who have no, we have to, who are not accountable to us, who are only accountable to shareholders, that's in the better case, or more likely private companies that are not accountable to anybody. If these tools are opaque, we don't know how they work. Again, they may, you know, as we saw with something like Grok a few months ago, hard to know in internet time, if that was a few weeks ago or a year ago, where you remember the period in which Grok was talking about, was sort of

[00:35:56] spouting white supremacy and, and, you know, it, the sort of, the tools themselves, and there's broader issues of power back to that in a moment, a question, which is, what can we do? What can I do? So it starts, number one, with some learning and education. And the great part about these, at least this new generation of AI, this generative AI, you know, Claude, ChatGPT, Gemini, these

[00:36:25] commercial tools, is that you do not need a computer science degree to use them. There is a version of all of these tools, which is free, which means I don't have to pay anything to try them out. I really do think we have to try them for ourselves in order to understand what they're good at and what their limits are. We have to try them and then start having the conversation about what can I do with

[00:36:51] these tools to actually organize my neighbors, communicate with my government, make the world a better place, which is a different conversation from, you know, I also like the conversation I have with ChatGPT around what should I have for dinner. I like the conversation we have where, you know, we upload pictures of my cat and make funny cartoons of my cat. Sure, that all that stuff is fun, but we have to know what these tools are. We have to use them a bit and we start to have

[00:37:21] asking them, how can we use them for our own community, for ourselves, for our family in ways that are good? So that personal education is extraordinarily important piece of this, but it's not at the expense of asking these bigger picture questions about regulating these companies, ensuring that the tools that we're buying, whether as individuals, as businesses, as schools, as governments, that we

[00:37:44] are either pushing back on these companies and making demands on them to say, we want to have tools that work. We want to have them be more transparent. We want to make sure they are tested. We want to make sure that before our governments start using AI to make decisions about our government benefits, that we know that how that's happening and how that works. So it has to start with some education.

[00:38:11] And then a lot of the things you talk about, we all are hearing about and talking about, which is regulation and accountability using our purchasing power, not just though to buy stuff, it's also to build stuff for ourselves. And let me, at the risk of going on too long, let me give you one quick additional example. We worked on a wonderful project with a group called Innovate Public Schools, a parent advocacy group out of California. My students helped to supply the engineering and AI talent to work with

[00:38:40] them and communities of parents to create a free, open source, non-commercial tool for building a tool that can help you understand your child's IEP, the individualized education program you get when your kid has a disability. It's about 15 to 20% of kids in public school get a document from your school district that says your kid is entitled to extra time on the test or some other kind of accommodation.

[00:39:08] That's a hundred page PDF in government speak. With the help of AI, those parents built a tool to say, translate that for me into my language, summarize it for me so I understand what my rights are and help me understand how I can advocate for my child with my school district. That's a free tool. We don't have to buy this stuff and we don't need these big companies with a little knowledge and a

[00:39:38] little bit of activism. We can go out and start making things that actually improve the world instead of making it worse. Yes. You know, the core of what I heard you say, thank you for that example, was really as citizens, we have to be involved in this process. So if in fact, AI can be used in all these different ways, we can't just be bystanders. That means, of course, we have to understand these tools, but also hold the companies, but also government accountable, right? And, you know,

[00:40:07] we use an example in a previous podcast. I was recently at the big TED conference and, you know, one of the things they're doing at X now, formerly Twitter, is using AI and their algorithm now to raise content where there's common ground and to flag content that is identified as misleading.

[00:40:32] When people from all sides of the political spectrum agree it's misleading, it raises that and reduces the virality by about 70% of misleading content. But again, we as citizens have to engage. We should be writing X and saying, do more of that and do less of something else, right? So we need to be aware of the ways it's being used in really constructive ways and where

[00:40:58] it's being used in destructive ways. We can't just, as you said, be just bystanders. Right. And the great part is that we can do that now because it's so much easier for us to learn how to use these tools. You know, 10 years ago, five years ago, you could have easily said, oh, that's for the engineers. That's for the computer scientists. I don't know. I'm the least, I just came from a meeting where two different people said, oh, I am the least tech savvy person in

[00:41:24] this room. That's a very common thing to hear. Oh, I don't know anything about tech. I can't possibly know anything about this. That's, we don't live in that world anymore. These tools, all you have to do is speak. You don't even have to speak English. You can just speak a language. You don't even have to type anymore. You can talk to control these tools. So there's no reason for us not to try them to understand what those limits are. So we can be better consumers. At the very least,

[00:41:52] it's about, you know, before you buy a toy for your child, that's going to start listening to and parroting back what your kid says, predicting what your kids should do, what social media platforms your kids are on. We now have just this better ability to gain insight into these tools that are coming into our bedrooms, into our homes, into our governments. Well, you know, our amazing producer, Jim Burke, who's in his mid seventies has about 10 AI agents.

[00:42:19] And he is like, he, you know, he is using AI squared. And so you're so right, you know, not to say he's hardly tech phobic, but it's a great example of how people are using it. I want to talk about one of your, I think, semi tongue in cheek suggestions in the book was that perhaps we come to a time when AI avatars would replace elected officials. So the collective,

[00:42:45] you know, will of the people, if you will, would create these avatars who would then make decisions. I assume you were at least mostly tongue in cheek, but I was thinking about years ago, I met a guy at the Hague from Argentina, the pre, pre large language models, but, you know, it was in maybe 2009 or something like that. And he, he was running to be the governor, he and his party to be the governor of the state of Buenos Aires. It's also the equivalent of a state,

[00:43:15] you know, not just the city. And his party was literally called the internet party. And their whole idea was ahead of its time in many ways is really was a plea for direct democracy. He said, we're going to use tools to consistently take the pulse of the people. And once we've given them as much information as we can to make good decisions, we're going to do what they ask us to do. But I

[00:43:41] was thinking about the founding fathers. Unfortunately, they were mostly fathers. Maybe, maybe we have even better constitution to more women were involved. That's just an aside. And I mean that quite seriously. But anyway, they did an okay, pretty decent job. But as you know, the founding fathers was, were as afraid of democracy as they were of tyranny. Yeah, because they didn't really trust the wisdom of the crowd. Talk a little bit about that. You know,

[00:44:07] how do you if this were to really evolve, right, into some form of more direct democracy, that isn't just a voting? Tell your thoughts on that. Would that be a good thing? Would it be a bad thing? What you know, and again, I the takeoff for me was your your comment about maybe eventually avatars will replace the guy be cheaper, by the way, to have the avatars take over from the elected officials.

[00:44:31] So I was deaf. Well, you know, it depends on the day of the week how tongue in cheek it was. I will say the part I was not saying facetiously was that was the idea that with let me go back one quick example. I had the great fortune last week to see a phenomenal demonstration of a project by the

[00:44:55] Scott Roberts Foundation, whose work I had not known previously working with ALS victims to use AI to give them back their voice. So as we know now, AI, you can clone somebody's voice. We know all the bad things about deep fakes, right? Sort of the fake in the political context, fake Biden calling and telling you, you know, not to go out and vote. Well, this very same technology they are using with a recording.

[00:45:22] But prior to the person having ALS, they can now give that person back their voice. And so I actually went to a presentation by somebody with ALS whose speech was very garbled and slurred and the AI helped them to actually communicate effectively and really rejoin the community of the living, if you will. So the not tongue in cheek idea was imagine if I could actually have a conversation with my political official in a way that and so, of course, the criticism comes of

[00:45:52] oh my gosh, you want me to talk to an avatar instead of to my politician. How often do you get to talk to your political official now? Right? So yes, in a perfect world, I would talk to my every, every constituent would be able to talk to their representative all the time. The reality is I never talked to that person. Even if I go to a town hall, maybe I get to raise my hand for 30 seconds, shout a question, get talked at and broadcast at. So the idea that I could actually communicate in a more

[00:46:20] direct way with my elected official, there are huge opportunities for reinvigorating like the representative. We don't even have representative democracy. Forget about direct democracy. It's not really representative. People with money have access on Capitol Hill. People who have lobbyists have access in the state capitol. You and I don't have that access. So I would just be happy if we could

[00:46:45] actually have the representative democracy that we were promised in any meaningful and real way, number one. Number two, I am personally not a great fan of purely direct democracy, what some people call sort of like push button voting, simply because I think the issues in our world are too complicated. Even something like how do I take out the trash or reduce the traffic? It's not about not having, giving direct voice to the people. It's that I can't just do it by going,

[00:47:15] yes, no, yes, no, yes, no. That's not how we take out the trash. We have to have a conversation about why can't we take out the trash? What's not working? How might we do it better? So I do think there are, let me call it more direct or what I would call more participatory ways we could do things, but it's not about push button voting and it never was. I do like the idea that a lot of people are

[00:47:39] proposing of reinvigorating the ancient notion of sortition, by which I mean in this case, the fact that in ancient Athens, political office rotated by lottery among ordinary people. And I just had somebody say to me yesterday, what if he really wishes that in his company or his school, and this doesn't have to be governments, it could be in our other organizations we're part of, that we each

[00:48:06] got to take a turn running the place. It doesn't have to be one person. Imagine if you had a committee that got to take turns being CEO, that got to take turns being principal, that got to take turns being part of the government. That's exactly what they've been doing outside of Brussels, where they have a committee of citizens that serves alongside elected officials, not instead of, but alongside of

[00:48:28] elected officials on parliamentary committees by lot. So there's a lot of things we might imagine, and it's not the alternative of just politician or somehow like push button voting. There are, I think, more robust, more participatory, and more realistic ways we could do things now that we can, frankly, organize that flow of information better. Sure.

[00:48:55] You know, what I like about what you're saying, and again, please get the book. It's a really fascinating, worthy read. Is that, yeah, there it is. And it's very attractive. It's quite pretty, I have to say. Yeah, I confess early on, like in the days when bookstores were, books were mostly sold, you know, in my early books, I sometimes would go and do that, turn my book face out in the bookstore. I confess I did it way back when. But one of the, I'd probably still do it today.

[00:49:23] Yeah. And one of the interesting things that you're really talking about is, is just the need for democracy to evolve. You know, as many people know, Jefferson thought there should be another constitutional convention every 20 years or so. Madison and other more thoughtful people said, we hardly got it done once. If we had to do it every 20 years, the whole thing might fall apart.

[00:49:48] But I think somewhere in between is what we need. So instead, we have a constitution written by people in a world entirely different than the one that we're in now. And a democracy that's hardly evolved at all in terms of the core way it functions. And I think what we need, what I love about your book is beyond just the AI tools, which are important in themselves, it's just we need to have a really

[00:50:12] robust conversation. I love that example you just gave about why can't the citizens be more involved, maybe direct democracy like that guy in Argentina had in mind isn't there, maybe because the founders were right, democracy is as worrisome as tyranny, or maybe not quite. But that somewhere in between, right? And I think it's an invitation to a more hybrid world than the either or world that we live in

[00:50:39] right now, right or left, the governed and the governing. You know, I just think that's what I think we need to be having more of those kind of right. I'm dismayed by the fact that you went to TED and there wasn't a conversation about AI and democracy. I'm upset that the UN has its big AI for good conversation. And there is no conversation there about democracy. So my hope is just that we start

[00:51:04] having these conversations about how to reimagine our democracy. That's why the conversation you're having here, episode after episode is so vitally important, because it opens up our just mental space, it creates those mental models for alternatives, alternatives that probably previously we were sort of just imagining, they were somewhat pipe dreams. But now we have tools that could allow us to actually

[00:51:32] reinvigorate to actually try these things in practice. And that's what's so exciting about I'm just so hopeful. Well, we have to be hopeful, there's no alternative. Well, and in fairness, there was a little bit of that conversation at TED, but most of it happened outside of the main stage, which says to me that, you know, we are really entranced by AI as a kind of

[00:52:01] technology, but not thinking yet about how it could really be transformative for democracy. And maybe also, my guess is, there's a fear around talking around this is what part of the problem, isn't it, Beth, is that in when society becomes so polarized, you made that point earlier, rigid is even a better word, isn't than polarized, right? Because one can be polarized, but not rigid. You know, you know, we had Monica Guzman on our podcast who talked about how do we have conversations with people

[00:52:31] who differ from us, you can, you can be polarized and not rigid. But in a rigid society, we're even rigid about democracy, aren't we? Instead of let's, let's, let's open it up and talk about how these tools could help us. And that's really what your book in this conversation is doing. Absolutely. You made the point about Ezra Klein before, and he powerfully makes the point in that book that we've always had people who've, they voted Republican, and they always vote Republican, they voted Democrat, and they always vote Democrat. And yes, there's some shifting, and there will be a

[00:53:00] generation long shift of who votes for which party. But we had for a variety of reasons, our ability to find common ground in between those polls. And we had, you know, and in our more complex world, with greater informational abundance, we simply need more robust and different institutions to manage everything from both how we take out the trash, to how we resolve our differences, that sort

[00:53:28] of essence of democracy itself. And that possibility, I think is here. Yeah. And you know, to your point, you know, when they called it up Capitol Hill, said we want to be more in touch with the people. But the good news for the, for the political leaders is if you really want it, that capacity is there now. And I'm guessing in almost all countries, there's more common ground than the political dialogue would suggest. So let's, Beth, time touches slightly, we're almost

[00:53:57] out of time where we could talk for hours. You mentioned hope. So I'm going to end with this question. What gives you hope? As you think about AI going forward and about democracy in general, I'll give you a shot at both. So what gives me hope is two things. One is I'll come back to shouting out some of these students, we call them our AI for impact fellows, they take a sabbatical from their studies, we do pay them,

[00:54:21] or I should say philanthropy pays them to be able to sit and build tools, like the team that has built a tool called GrantWell for the state of Massachusetts to help small towns and communities apply for public money to do things like replace their old school buses or take the lead out of their pipes. Those students then not only have built that tool for free, given it to Massachusetts, but now are giving it to and building it for California, Rhode Island, and a bunch of other

[00:54:50] states so the taxpayers don't have to pay and we can and more communities can get access to that public funding. These kids get a job out of it. And they get to do meaningful good work in the world. I'm also super excited one more shout out to the peer to peer learning community that's also happening online. We run an online community called Innovate US where people can come for free, it is no cost, it is never any cost, again, supported by philanthropy to learn about how do we use AI for public good.

[00:55:19] It is peer to peer people teaching one another about what works. It is the team that built that GrantWell project teaching other people, how did you build it? What did you do? What could we do? So it's creating that conversation among people who are want to use these tools and other tools, frankly, to make the world a better place and especially to repair our own communities in these troubled times. So I am hopeful.

[00:55:45] Excellent. Well, Beth, first of all, thank you for being with us today. Thank you for your good work. This is the book, Reboot AI and the Race to Save Democracy. It's a worthy read. Thank you. Just in closing, a couple of thoughts for our listeners. You know, Bill Gates was asked about AI's future and he said something very similar to what Beth said a few minutes ago. He said,

[00:56:09] every human technology we've ever invented could be used for good or evil. And so we need to get out of this binary world that AI will either save the world or destroy the world. The truth is, it is an amazing tool that will transform society. But how it is transformed will really be up to us. And especially when it comes to democracy, it gives us the opportunity to bring people together, to find common ground,

[00:56:36] to change the relationship between the governed and the governing from pretty much a one-way and once-in-a- few-year-old enterprise into an organic community, together figuring out what the future might be. And as Beth said, it's really going to be up to us as citizens which way that goes. We can't passively stand by and hope for the best. We have to fight for the best. As you know, on this podcast, we always

[00:57:03] remind you, history is not a destination waiting for us to arrive. What we will one day call history will be the result of what we do today. So we'll see you next time on The Way Forward. Good day. Thank you for joining us in this episode of The Way Forward Regenerative Conversations podcast. I'm Jim Burke, the producer. If you found this conversation as enlightening and motivating as

[00:57:30] we did, please subscribe to our podcast and YouTube channels. Your subscription helps us continue bringing these vital discussions to a broader audience. Don't forget to share this episode with your network and let's work together to build a better world for all generations. Find us on YouTube, Apple Podcasts, Amazon, and now on Instagram and YouTube Shorts. Together,

[00:57:56] we can make a difference. We'd love to hear your feedback on today's episode and any suggestions you have for future guests. Please send your thoughts and recommendations to thewayforwardrc at gmail.com. Your input is invaluable in helping us shape the future of this podcast. Thank you for being part of the Way Forward community.