Denizen

Deliberative Democracy and AI with Claudia Chwalisz

Episode Summary

How can artificial intelligence help us scale deliberative forms of democracy that are more resilient to autocracy? What aspects of the deliberative process are fundamentally human? What are the risks associated with using technology to scale innovative democratic process?

Episode Notes

DemocracyNext founder and CEO Claudia Chwalisz returns to the Denizen podcast to explore the intersection of deliberative democracy and artificial intelligence.  Here we are particularly interested in how we might crease democratic resilience by leveraging AI to scale deliberative processes.  Governance innovation is the AI application we're most enthusiastic about so we're excited to bring this conversation to the Denizen audience.

Claudia and Jenny discuss broader trends of rising autocracy, weakening civil societies, and attacks on basic rights, while situating deliberative democracy as a counterforce to these issues.  They explore what aspects of the deliberative process are fundamentally human, giving rise to the limitations and risks of certain applications of AI.  They also break down five different ways we can think about scaling deliberative democracy, which enables a more nuanced conversation about the opportunities AI presents.  Claudia stresses that technology is not a silver bullet, but must be integrated with the human elements of social and political change.

 


00:00 Introduction to the Episode

00:32 Meet Claudia Swi: Expert on Deliberative Democracy

00:43 Exploring Deliberative Democracy and AI

00:55 The Importance of Democratic Resilience

03:03 Understanding Populism and Authoritarianism

04:49 Challenges and Limitations of Representative Democracy

08:26 Defining Deliberative Democracy

09:49 The Role of Sortition in Deliberative Processes

13:54 In-Person vs. Virtual Deliberative Processes

23:13 The Human Element in Deliberative Democracy

31:23 How AI Can Support Deliberative Structures

35:18 Exploring Tech-Enabled Trackers and Feedback Loops

35:56 Understanding the Lifecycle of Deliberative Processes

36:27 Defining and Breaking Down 'Scale' in AI and Democracy

37:08 The Five Dimensions of Scaling in Deliberative Democracy

37:19 Scaling Out: Increasing Participation in Deliberative Processes

40:29 AI's Role in Scaling Out and Up

46:24 Scaling Across: Spreading Deliberative Processes

49:37 Scaling Deep: Institutionalizing Deliberative Processes

52:45 Scaling In: Enhancing Quality and Impact

59:56 The Broader Implications of AI and Deliberative Democracy

01:06:03 Concluding Thoughts and Future Directions

 

Resources:

Episode Transcription

[INTRO]

Claudia Chwalisz: [00:00:00] But at the end of the day, none of these issues are purely technical. They are grounded in values and what kind of society do we want to live in? And to be able to answer those questions, we do need some purely factual objective information, but we also need to have spaces for people to hear each other and to.

Decide, okay, the science tells us this, but do we want that to happen or do we want something else to happen? So that's a very human part of the discussion as well.

Jenny Stefanotti: That's Claudia Chwalisz, it's founder and CEO of Democracy. Next, a global platform for democratic innovation. This is the Denizen podcast. I'm your host and curator, Jenny Stefanotti. In this episode, we're talking about scaling deliberative democracy and the role that AI can play. Democracy and AI are the predominant topics on most of our minds these days.

So I'm excited to bring you a conversation intersecting the two. What we're really after in this episode though, is exploring how we can use deliberative processes [00:01:00] and technology to cultivate democratic resilience that is democratic governance structures that are resistant to the forces of autocracy.

You might recognize Claudia's name. She's been a guest on the Denizen podcast before last year. We covered citizens assemblies with her, which is a pretty central topic to understand. So if you haven't listened to that episode, I'd highly recommend you check it out as well. This conversation builds on that one, exploring how we can scale the deliberative processes like citizens assemblies, particularly with the use of new AI enabled technologies, which is the most exciting use of AI in my opinion.

Claudia is an incredible expert on deliberative processes and populism. Her work is at the forefront of democratic innovation, so we're super lucky to be learning from her in addition to founding Democracy. Next, she's also the author of several books, including The Populous Signal, why Politics and Democracy Need To Change, and The People's Verdict, adding Informed Citizen Voices to Public Decision Making.

As always, you can find show notes and the transcript for this episode on our website becoming [00:02:00] denizen.com. There you can also sign up for our newsletter. I bring our latest content to your inbox alongside information about online and virtual denizen events, in addition to announcements from our many partner organizations.

All right. Without further ado, here's Claudia. 

 

[INTERVIEW]

Alright, Claudia, it's great to have you back. It's great to be here. So Claudia has just released a paper on deliberative democracy in ai. AI applied to democratic innovation is the use of ai. That's always gotten me excited, so I'm very looking forward to double clicking on what that might look like in this conversation.

So thanks for your time.

Claudia Chwalisz: Great. Well, I'm looking forward to the questions you're gonna ask as well.

Jenny Stefanotti: Well, I,

Claudia Chwalisz: I just

Jenny Stefanotti: wanted to just start with some context setting. Can you put the work that you're doing in terms of deliberative democracy? And we'll talk about democratic resilience in a second, just in terms of overarching trends.

Towards autocracy, weakening civil society attacks on basic rights and freedom. Anything else you wanna say about the big picture moment that we're in?

Claudia Chwalisz: I think the things you just said sort of capture it as well. [00:03:00] At the same time. I think that it has been building up for a while. Like this is not somehow just in the past year either, that this has new or changed my own entry into doing this started about 15 years ago with research on populism and on authoritarianism, and wanting to understand the extent to which people's disillusionment, their lack of agency was driving people towards these different populous parties, actors, and becoming really convinced from that time onwards, it's that this is part of the deep trends that we need to address.

So I think it's only been accelerating over the past year and it's definitely through that political lens that I'm doing this work when I'm thinking about AI and deliberative democracy and citizens assemblies and other forms that it takes.

Jenny Stefanotti: Thank you for reminding me that that was the genesis of this work, that inquiry for you.

Is there anything on that note before we move on, just anything that you feel like the audience should know about this trend towards populism and what's creating the acceleration in recent years?

Claudia Chwalisz: I don't have all the answers on that either, but it's also a complex phenomenon. So what I was [00:04:00] mentioning just now is that I was really interested in better understanding this, let's say, political or democratic dimension, because when I started this work about 15 years ago, people were really focused on the economic drivers, the cultural drivers of these trends.

And it's complex. So those are part of it as well, of course. But if we recognize that there is this dimension of people feeling, lack of agency, lack of voice, disillusionment with the political system, then you don't just think, oh, we just need better economic policies. You recognize that there's also some need to change the political system to give people a meaningful say in shaping decisions affecting their lives.

And I think when we look at a systemic level, I don't think people have had a really big shift in power and in agency. And if anything, those trends have actually been getting worse, as you were saying in the introduction about curtailing civic freedoms and otherwise.

Jenny Stefanotti: I think it's also really interesting to look at it in the context of just fundamental limitations of representative democracy and representative democracy in the context of capitalism and [00:05:00] neoliberalism in particular, and the way that those political and economic systems interrelate.

We talked about this in the Citizens Assemblies conversation, the limitations of the four year election cycle, why four years is well too long and too short, too long to have good feedback loops and correction mechanism, and too short for incentives to be long-term oriented for policy makers.

Claudia Chwalisz: And that it's by design as you're saying.

Jenny Stefanotti: Yeah, so it's just interesting in the kind of like the feedback loops that are present in the system in terms of corruption and that being an accelerating factor with the trends that we're seeing. So I really appreciate the conversations about deliberative democracy in the context of the broader landscape from a theoretical perspective of representation and delegation and direct democracy, and the ways in which these models address some of those fundamental issues with representative structures.

Claudia Chwalisz: Indeed. And it's not about getting rid of representation, but it taking a different form and also power rotating in different ways.

Jenny Stefanotti: Totally. Your paper starts talking about why deliberative democracy is [00:06:00] interesting in terms of democratic resiliency. So let's talk about that. What do you mean by democratic resiliency and why don't you think that we have it in the current institutions today?

Claudia Chwalisz: It felt really important to actually, to put that at the very beginning, opening part of this paper because I find that a lot of the conversations that have been taking place about the intersection between technology and democracy or deliberation in particular of what we're talking about here have a starting point that's often maybe a little bit more technical.

You start focusing right away on like increasing numbers or different aspects of this and, and there's an importance to kind of take that step back and think like, hold on. Why are we even interested in potentially doing that in the first place? And the main motivation for me is rooted in wanting to have more meaningful ways for people to have a sense of agency.

And have a society where a citizenry has a greater sense of being empowered, has a voice, has a way to influence decisions that are affecting their lives. And I think that those deeper roots [00:07:00] also of the people who have that sense of connection and belonging and trust are interlinked with that. So at the end of the day, for me, my interest in how do we scale deliberation, deliberative processes is linked to how do we address these deep democratic problems we have, so we have a more democratically resilient society.

Jenny Stefanotti: Yeah, I appreciate that. Acknowledging that there are a lot of people in the space who are more kind of hammers looking for nails versus from first principles. What's the problem that we're trying to solve? And your paper outline specifically that the elements of democratic resiliency are that people feel heard, people empathize with one another.

Citizens have strong interpersonal and societal trust, and that public decisions have legitimacy and alignment with collective values. So it's interesting to think about also the context that we're in now where there's more polarization so people don't feel heard and they're not empathizing with one another.

Trust is declining across the board. There's a lack of legitimacy [00:08:00] because public decisions don't reflect public values. They reflect corporate interest in many cases. So I appreciate the way that you broke that down in the context of the existing environment and how these different structures shift that.

Okay, so let's hold the frame of democratic resiliency and how these new structures and AI can can help foster that. We haven't done a 1 0 1 on deliberative democracy, and I appreciate it again that your paper started there too. So let's just make sure that the audience understands like, what is deliberative democracy mean?

Claudia Chwalisz: I think that's a good point because not everyone is super familiar with this, and I think even if you've studied political science or so on, we tend to think of democracy as elections and deliberative. Democracy is, I think, a fundamentally different mindset when we're thinking about what does democracy entail, and we'll go down the route of too much of the deep political theory, but the basic premise is that we should have a democracy that's rooted in people.

Exchanging reasoned forms of arguments with one another to get to some sense of [00:09:00] collective values, collective decisions off the back of those, to have those spaces where we are able to have those forms of exchange, and it's on that basis that we're taking collective decisions together with one another.

And then there's different, more specific forms that deliberative democracy can take. So in the paper, we also are a little bit more specific to make this something that actually has a, I guess having a little bit of boundaries about what are we talking about as well. So I think one of the main forms that deliberative democracy as a wider, broader political theory takes in practice is in the form of what in the academic literature get called deliberative mini Publix.

When I was at the OECD, they were called the representative deliberative processes. I think most people, if they've heard of these ideas, would come across notions of like citizens assemblies or citizens juries or civic panels. So these sort of forums where you have the main criteria of people who are part of them are selected by sortition, which means random selection of decision makers.[00:10:00]

So it sounds like a kind of technical wonky term, but it's actually just a very ancient concept that's being revived. So sortition, and then having structured spaces for deliberation, which means collectively weighing evidence together, working on the basis of that to be also able to, well, also sharing their collective values and listening to one another, sharing perspectives, and doing the hard work of finding common ground on their shared recommendations.

So in these many publics, usually there's a threshold of consensus of around 75, 80% that needs to be reached, for example.

Jenny Stefanotti: Yeah.

Claudia Chwalisz: So just to make that wider theory feel more concrete, how we live out deliberative democracy tends to be in some of these practical forms that have some structure around them, and therefore enables us to think through some of these framework elements around dimensions of how do we scale this if we're interested in doing that.

Jenny Stefanotti: Yeah. And so I think just to underscore that, right, we're talking about. Spaces where people with different perspectives can come together [00:11:00] and learn about the issue, share their perspectives on the issue, learn from each other, and come to consensus on policies around those issues where the policies have implications for the people's lives who are represented by that body.

And in these mini publics, we're talking about this sortation process where it represents the public, but through ordinary citizens rather than elected officials. And there actually is real representation because of the sortation process, not theoretical representation, but then through the actual process, somebody may not be represented at all.

Because if you have a two party system and your party loses, you're just screwed for that particular election cycle, right? And so that the fundamental difference is there is actually guaranteed representation and it is actually ordinary citizens. And so you don't have all of the complex corruption effects of the default political systems and the way the money is corrupted politics as we know it.

And then we're also talking about these structures alongside the larger political institutions [00:12:00] as we know it.

Claudia Chwalisz: Because you have more of a rotation of power when everyday people are randomly selected for one assembly. Then there's another assembly with different people who are there. And I mean, I don't think we go into this level of detail in the paper, but even the roots of democracy, the principles of sortition within in ancient Athens, many people don't really realize that actually, when we often make reference to this model, but in Athens there were 7,000 posts and only 100 of them were chosen by election, and the rest were through sortition.

So it's a practice that was so much more common. And we've sort of lost it and lost even the notion of like that that's actually the democratic form of selection. Whereas elections have existed. But those were always considered to be an oligarchic form of how we choose people. Like even the etymology of the word election is with the same roots as the word elite.

It's a form of how we choose our elites rather than how we choose the people. So citizens, assemblies, deliberative, many [00:13:00] publics are a sort of modern incarnation of some of these really ancient democratic ways of decision making. Hmm.

Jenny Stefanotti: I appreciate that. And I think another critical point is, and we talk about deliberative democracy, we're not strictly talking about the public sector.

We're talking about these kind of decision making practices writ large in society where decisions have implications for people's lives.

Claudia Chwalisz: For sure because these principles of sortition and deliberation can be applied in other organizations. So for example, at Democracy Next, the organization that I'm running, we've been doing work with some museums with Mongan Corporation, which is one of the world's biggest and oldest cooperatives.

So it, it really shows that actually there's other types of organizations that can be rethinking how their governance structures work, but also how they're taking decisions on a kind of more daily decision making cycles on an ongoing basis as well.

Jenny Stefanotti: Yeah. I just wanna make sure people were clear. This doesn't necessarily apply just to the public sector.

I think one interesting point about deliberative mini publics is that it can [00:14:00] be in person or virtual or hybrid.

Claudia Chwalisz: Actually, that's a fair point to make sure that we make explicit for the sake of the conversation. When I'm talking about deliberative mini Publix, I am talking about processes that are almost always, most of the time, at least most of the examples we have, and, and the ones I'm most interested in personally as well, are in person because we are interested, again, going back to what problem are we trying to solve.

I don't think there's a magic shortcut to people feeling greater sense of trust, belonging, connection, less polarization than spending lengthy periods of time together listening to each other, speaking, hearing each other, doing the hard work of finding common ground and doing something. So most of these assemblies are in person.

There are a number of examples of these taking place fully. Virtually or in some hybrid mode as well. But COVID also showed us, I think some of the challenges of when you're trying to apply these principles of taking a process that has been really done largely and designed largely to be in person, to be [00:15:00] fully virtual, you need to really rethink it.

And I think certain aspects of it are just more challenging. I think anyone who's also worked on online virtual, global community building or other thing knows this because, for example, in citizens assemblies, one of the most important aspects that enables people's ability to grapple with complexity during a deliberation is what gets called group building in the academic literature, which refers mostly to the very informal things that happen at the beginning.

Of people having coffee, having time maybe to have dinner the night before meeting, getting to know each other, having time to talk about their values before they actually dive into some of the deep material or stuff that's controversial or things they're learning about. And again, that part when people are designing virtual things either gets skipped entirely or it takes such a minimalist format that you don't necessarily get the same benefits for it.

And I think also one of the other things that, again, thinking what problem are we trying to solve? Democratic resilience and making these processes genuinely democratic and not [00:16:00] just technocratic in nature. So in assemblies, people tend to get paid for their time to be able to participate. Also childcare is provided and there's all these measures made to make it equal.

But in the virtual processes with sortition, we have found in, in numerous examples, like there will be, depending on the place between fifth and a third of the people who don't have access to have a computer at home that they could do all of the sessions on. Because maybe there is a computer, but they have a family that they share it with.

Maybe they don't have it at all. Like we spend a lot of our time on Zoom and on conference calls, but a lot of people don't. So there's also such a learning curve when you're actually wanting to involve a wider public that is genuinely broadly representative and not just the volunteers who are tech savvy, who know how to use this stuff.

So I think sometimes there's this assumption that, well, if we just make it fully virtual, it'll be like easier or cheaper to organize. But actually if you're committed to the Democratic principles, which for me really count [00:17:00] because of why I'm doing this work in the first place, it's actually harder and I would say even more expensive to do it that way.

So I think there's trade offs involved and I think the hybrid approach has worked a little bit sometimes. And if there's in between sessions, 'cause you can also provide people with materials or do a bit of training on how to go on Zoom or stuff in like when you're in person before people connect online and things like that's actually really necessary.

Jenny Stefanotti: You're making two really critical points. One, from an accessibility, truly democratic perspective, going virtual is suboptimal. It can actually be much more clunky and people just can't access it. And secondarily, there is no substitute for being in person for really building relationships and having a foundation of compassion and curiosity for the deliberative process.

And there were so many amazing stories of thinking, of course, immediately to that story of Ireland and the same sex marriage or those two people who would've never listened to each other, just connected to each other as humans over beers. And he teed up a deliberative process of coming to [00:18:00] consensus on a very contentious issue.

And

Claudia Chwalisz: I mean, that took five months to get there, right? Yeah. Like that didn't happen immediately the second they met either. It takes time to build trust with people. So I think sometimes with this period of trying to make things more efficient, we actually lose more than we gain on other. I guess facets we're interested

Jenny Stefanotti: in.

Yeah. And then can you just say a little bit about Async versus sync?

Claudia Chwalisz: Yeah, that's a fair point actually. So again, yeah, I'm talking about synchronous processes because part of deliberative democracy actually entails people having conversation and being in dialogue and kind of really back and forth with one another.

Exchanging arguments, exchanging points of view. And with the way that a lot of these asynchronous tools are designed, it's not really a conversation. And that's not to say that tools like PS are not useful and that they can't be used for different purposes, or that they can't be combined. And actually.

Be enhancing as citizens assembly or other deliberative process, either before or during or even after. There's different ways this has actually been [00:19:00] used and designed, but one thing we do make clear is that for us, and I say us, 'cause this is a paper I co-wrote with my co-author, Sammy McKinney, we don't actually think that this is deliberative or fits the definition of deliberative democracy and or deliberative mini publics, at least in the way we're talking about it as well.

So I think it's important, again, to clarify the boundaries of what we're saying. And again, I don't wanna come across as disparaging of those tools either, but I just think we need to be clear about what we're talking about and how it's different.

Jenny Stefanotti: A hundred percent. No, I really appreciated that distinction and, and actually I'm finding that really thinking deeply about what requires in person.

What can we do uniquely in person? I'm fielding something like this right now with the Denizen Council, where there's a lot of us and there's a lot to talk about, and we don't have a lot of time together. And if we just use the time together to have the entire conversation where we're surfacing everybody's point of view and trying to look at where there's synthesis and alignment and where there's a distinction and we just simply don't have enough time.

And so what I've actually started to do is async get [00:20:00] everyone's opinions out and do the synthesis to find the sources for the discourse in person. Then that means 60% of the process actually can happen async to tee up what has to happen, sync, and it's really improve the efficiency of our processes.

Claudia Chwalisz: That's a nice example of actually demonstrating how these things can be used.

Jenny Stefanotti: Yeah.

Claudia Chwalisz: Together. And I think from what you're describing as well, you can see how using that asynchronous tool in itself by itself is not deliberative totally. But when you combine it with another kind of process, it becomes something that enhances a deliberation.

So that was the only kind of distinction we were making.

Jenny Stefanotti: Yeah, really appreciate that. Let's just briefly talk about scale. What are the inherent ways in which absent ai, there's some scalability issues with these deliberative structures?

Claudia Chwalisz: Yeah, because there's a trade-off in terms of the quality of a deliberation you can have and the number of people that are involved in any one deliberative process.

And I think anyone who's been part of a [00:21:00] larger convening where maybe there were even a hundred people, and you're trying to get to some shared consensus on something at the end, you can see how hard that already becomes. So a lot of these deliberative mini publics. Obviously, depending a little bit also on what scale they're happening.

Like when it comes to, is it local government and national government, you're trying to get some legitimacy and sense of wider representativeness. So some of the national level ones have been in that 150 to 200 people range, which in some ways is also still not massive either. A lot of the local ones tend to be more like 30 to 50 people, which in my view is actually a quite optimal size for a high quality deliberation where everyone knows one another has had a chance to build trust.

Everyone can actually intervene properly in a plenary context. You resort a lot less often to voting. You use it at key moments to really make sure you're not getting some false consensus either. But if you introduce voting too early into a process, you also have other knock on effects that are not necessarily optimal for the quality of a deliberation too.

Right. So [00:22:00] that is just. An inherent trade off when it comes to scale within these kinds of processes. And then I think another dimension that has nothing to do with technology has to do with, well, I guess two of the other aspects we talk about in the paper, like one is how do we just make this happen more often in more places?

Which relates to more people even knowing about these ideas and having the skills and competencies to know how to do that. That's part of the challenge of scaling right now too. And then there's another challenge. Which relates more to where I think the relational aspect is most important, which has to do with ensuring that these processes really have the highest impact possible.

That the people in them are as empowered as possible within a certain context. Getting these to not just be one-off initiatives, but actually starting to become institutionalized or made permanent, like with legislation that anchors their existence and their relationship with an elected body and a public administration.

Doing that work of having [00:23:00] many conversations with all the elected officials, public servants and so on, like, and AI is just never going to do that effectively.

Jenny Stefanotti: Yeah. This is organically bridging us into the next question, so let's integrate them. What about this is fundamentally human. So part of this is we're talking about inherent limits to scale.

And I think this gets into this question of how much can AI support these processes inherent limits to scale around how many people can reasonably actually have a conversation that is, by definition deliberative. It becomes more like opinion polling and aggregating as the number goes up, you simply cannot have a collective conversation.

And then I think another thing that you pointed to, which is really critical is relationships and trust is really important. And once N goes above, we're all familiar with Dunbar's number. The ability to hold those relationships is is much harder. So that's some things that are inherently not scalable about these processes, but let's talk about what is inherently human that can't be replaced, maybe can be enhanced or supplemented or [00:24:00] made more efficient, but what is inherently human that can't be replaced by machines within these processes?

Claudia Chwalisz: Yeah, it's a good question. I think it's also an interesting framing. 'cause even the way you've asked it has really actually made me think of like. Humanness versus maybe other qualities that I've been interested in or are thinking about because I guess the parts of why I'm attached perhaps to things happening in person and synchronously, I think relate to this human nature of it, which means that, I guess in some ways you are creating a space where for this to be a very effective process, you do want people to build trust, but therefore also be able to show some sense of vulnerability with one another.

To open up in that way. Like you referenced this example from Ireland, which. I mean, if people are not familiar about this friendship between an older truck driver who has been openly homophobic for most of his life, and a younger tech guy from Dublin in the city who made a point of really wanting to his, his gay [00:25:00] identity to be very obvious, and they end up sitting together at the table where they were randomly assigned on the first day, and over the course of five months, become super close friends, which they are to this day.

I actually just learned there's a documentary being made about their story, but the fact that they were able to open up to one another, it didn't happen just in the assembly, but also the moments like of going to the pub after, well, the pub might be very British, but you know, going for a kombucha in some places or whatever you might do elsewhere in the world.

Like those things are kind of fundamentally human and also not peripheral to what makes these processes really work.

Jenny Stefanotti: Yeah, I think it's really interesting. The paper spoke to the deliberative process and the ways in which values are integrated. So I think the presence of values. Is something that is inherently human.

I guess you could argue that you could embed different representative LLMs that might be able to have those values and have a deliberation around those [00:26:00] values. I'm sure there are

Claudia Chwalisz: people who are arguing that are trying to do that. No, but precisely, I think actually that's a very good point that you bring out and tease out because I think having time and space for that conversation about values and what are our shared values in relation to this topic that we're deliberating on is in itself a really crucial part of this process, and as you say, is also human.

And I think we are going down a very dangerous route if we're trying to skip or eliminate that or make assumptions based on calculated preferences by LLMs about such things as well.

Jenny Stefanotti: Yeah, I mean it's so interesting, like the whole deliberative process fundamentally is trying to consider a heterogeneous set of values held across society or the people affected by decision making to come to the effective set of compromises to make policy decisions.

So I think that kind of representation of the values is something that is interestingly and inherently human and inherently cultural. But you know, another really interesting thread of the Tenon inquiry is we talk about from a cultural perspective, [00:27:00] this legacy of the enlightenment and the scientific revolution.

That the overweights reason and, and intellectualization and rational thinking and the need to be more embodied and understand the wisdom in our bodies. And your paper also speaks to feelings. Feelings are also inherently human and deliberative processes that help us feel our feelings and understand the wisdom of our feelings and bring that into deliberative process, I think also is fundamentally human.

Claudia Chwalisz: Yeah, and actually we, we, we emphasize that in the definition around deliberation. I actually earlier stressed reason, but you're right that I actually, maybe I also mentioned emotions, but I do think it's really important to recognize that it's not just an intellectual exercise, that there is also a need to have this space for people to be grappling with.

And listening to each other is like, well, listening first to themselves, but also to each other is expressions around the feelings that come up around the issues they're deliberating on. Because it's precisely why you need to have spaces of deliberation around [00:28:00] issues that some people might be dismissing as purely technical.

So we just need expertise on them. But at the end of the day, none of these issues are purely technical. They are grounded in values and what kind of society do we want to live in? And to be able to answer those questions, we do need some purely factual objective information, but we also need to have spaces for people to hear each other and to.

Decide, okay, the science tells us this, but do we want that to happen or do we want something else to happen? So that's a very human kind of part of the discussion as well.

Jenny Stefanotti: Yeah, yeah. The kind of embodiment really sitting with how does this feel in my body? And I'm just very interested in decision making processes that really integrate that embodiment component of it and, and something that we've been exploring in these conversations.

Mm-hmm. So just think that's something to weave into what is fundamentally human that can't be replaced by ais and how AIS can be supportive. And I know that there's a lot of debate, I'm sure in the field of, well, why couldn't we just have an LLM that represents this subset of the population and [00:29:00] something else?

And we talked about an LLM that's trained on nature and represents nature and governance. So it is interesting to think about AI agents representing right in some of the discourse. But I think we know what we're touching on is there's still something fundamentally human about values and feelings and emotions and embodiment that cannot be replaced.

For sure, and I think really critically, your paper makes some very important points around all of this. Technical innovation is interesting, but it has to be in the context of also cultural and social and process design innovation as well. It's not strictly a technical problem. It's a design problem around the entire process and how to scale it.

Claudia Chwalisz: For sure. And if we go back to what problem are we trying to solve, there will always be a relational human element that's needed to achieve the scaling goals if we keep that in mind. So I think really we try to foreground that part of the argument because I find in some of the discussions that have been happening around these questions of how do we scale this?

That part usually gets [00:30:00] rather neglected and there's a jumping to claims about, well, we could just make this like easier and cheaper by eliminating actually the human facilitators or moderators that are also, I think another part that's really essential because when you are creating a space that people are talking about controversial issue, like same-sex marriage, abortion, or euthanasia like in France, that just happened.

You also want a space that you have really skilled, trained facilitators that are there with people having those conversations too. So the process in itself matters. Sometimes there's this sense of we can be so much more efficient if we use technology to perhaps collapse a part of that back and forth, and you can get, I think, to the same decision outcome.

But if you eliminate the process of all the people involved in being affected by or implementing that decision, actually having had the conversation with one another around it, you don't at all have that same sense of ownership or agency or buy-in around it. So. It matters for a few different kind of [00:31:00] reasons.

Jenny Stefanotti: Yeah, I was about to make that point too. I just really appreciate like, part of the value of these deliberative mini publics is that citizens have an embodied experience of contributing, and that in and itself changes trust in public institutions. You have to actually be in it, in your body as a human being, which means don't have, and I'm sure there's a lot of really interesting nuance debate happening in the field.

We're not gonna get into the weeds there, but I appreciate getting into that. Okay. So what are the ways in which AI can support these structures? Let's look at the lifecycle of setting up and running a mini public and where AI can support it, in particularly, I'm interested to hear where are you most excited about ai?

Claudia Chwalisz: Yeah, maybe to break it down first for people who might be less familiar and so we're all kind of thinking about or visualizing maybe the same thing. So I think the first high level way you could break it down is everything that happens before an assembly even meets in the first place. Everything during the deliberation itself, and then everything that happens after when it comes to a link to implementation of recommendations and [00:32:00] follow up and feedback loops and everything.

And then within

Jenny Stefanotti: each of those, quick question, with the deliberative process itself, would you break that down into getting smart on the topic and then deliberating on the topic? Or is that just such an integrated thing that it's all one? Because I know there's a lot of presenting the information and the context and the experts and getting more testimony, and then is it delivered or is it sequential or is it just really like you're constantly learning, talking, learning more and talking.

Claudia Chwalisz: At a certain point, it stops being as sequential. I think like if we zoom into that, like what's happening during an assembly process, you actually tend to have designed in before even any of the learning happens, this phase around group building or sharing of values and defining those for the group. And then you get into you usually some wider learning.

That happens first, but the more people already start to learn, the more the deliberation and the learning melds together because people are, have already a more informed basis on which they're learning new things, asking other questions, and so on.

Jenny Stefanotti: Nia, there's just an informational baseline aspect of the [00:33:00] deliberate process where AI can help a lot by synthesis and presenting a lot of information very succinctly to people versus the actual discourse, which there's more limitations.

Right. I just wanted to distinguish between those parts of that second part of the process.

Claudia Chwalisz: Right. But I also think that this is one thing where, again, I think we're in an experimental phase of what we're doing within assemblies and what's optimal and so on. 'cause there's also a whole literature in the field of education and pedagogy, which also shows that there's also precisely the value of a process of people needing to do some hard work, to learn something that people actually learn best when they have to present information that they've learned back to other people.

That actually when people write things by hand versus when they type them, they retain that information and that knowledge better. So again, we also need to take into account, yes, AI might be helping with some aspects. And I definitely think there are many things that are exciting, but I think that constant sort of questioning of like, okay, this is possible, but actually is this [00:34:00] desirable, is something that we really need to maintain as we start to introduce more and more like different kinds of technology applications within these deliberative processes as well.

Jenny Stefanotti: So again, the three parts that you talked about is setting it up, running it, and then what happens afterwards. In a general sense, where in that are you particularly enthusiastic about the use of ai?

Claudia Chwalisz: What I've been most involved in actually testing out has related to this stuff like that's during the assembly process.

We've been doing some collaborative work with the MIT Center for Constructive communication around how we could be using different tech enhancements during an assembly process. And we've only, I think, just scratched the tip of the iceberg of all the ideas and things we've been floating around, and I know others in the fields are also thinking about this.

I, I do think that, I guess when we're thinking about some of the other dimensions of what would enable maybe the spread of these assemblies rather than scale to better convey the notion of how do we make more of these happen more easily? I think having better technology for that whole before phase, [00:35:00] which I think it's a little bit neglected, but there's often that least four to five, if not six months of work before any deliberative process happens to set it up to prepare it.

And of course some of that will always remain relational work, but there's also a lot of work that's rather repetitive in nature where I think technology could be utilized in much better ways. And it's not always actually with some of the latest L LMS that we could be doing stuff too, even though I think there's some applications for new kind of tools that can be built much more easily today to help the field.

And then I think, again, with an interest on impact and ensuring that something actually happens with recommendations that come out of this. I am interested in thinking about how we can build better tech enabled trackers, but also feedback loops and things around recommendations that have come out of different assemblies and and so on.

So I'm thinking rather holistically actually about all of these different phases and we can kind of talk in a little bit more detail and stuff you wanna talk about.

Jenny Stefanotti: Yeah, let's get into it. But I just wanna make sure that just the audience has in their mind as we're talking about [00:36:00] this, the life cycle of this, which is setting it up the actual deliberative process, which again includes the group building relationship building, getting smart on the topic, deliberating on the topic, policy recommendations, and then what happens on the backend in terms of implementation and communication to the wider public.

Because as we talk about the different ways that we might scale, there's relevancy for different parts of that lifecycle for ai. So now let's get into it for sure. Because what I really appreciate about your paper is that AI is such a broad thing, right? And scale can mean so many different things as we'll see in momentarily in the context of deliberative democracy.

And so the really meaningful contribution of your paper is to say, okay, what do we mean by scale? Let's break it into five different ways that we can talk about scaling these structures and where within each of those five ways to scale is ai. Interesting. So that's what we're going to get into now. So the big contribution of your paper to really just give us some framework and scaffolding to have a much more [00:37:00] nuanced conversation about.

What the opportunities are, what the limitations are, and what the risks are of using AI in this space, which is so exciting. Hmm. Okay. Drum wall now for the five dimensions of scaling. This is the things that we get excited about on this podcast. Okay? Yes. Alright. Let's talk about scaling out. That's the first one that you talk about, which I think is the thing that people generally think about when they think about scaling.

So what does scaling out mean?

Claudia Chwalisz: I think this will be the thing that maybe intuitively, most people just assume everyone will understand if they just say scale. But it, it refers to wanting to increase the number of people in any one deliberative process.

Jenny Stefanotti: Yeah, but I appreciate that. You mean not just directly?

Although you mean directly, right? So it could be 30, it could be 150, it could be hundreds.

Claudia Chwalisz: True. We, we distinguish actually between two dimensions.

Jenny Stefanotti: Yeah.

Claudia Chwalisz: Yeah. Also that I think most people are thinking about it in that direct way. I think within the wider [00:38:00] field there's also a lot of people though interested in how could we, at least in some way or other, involve a wider group of the public that either contributes to or engages with the smaller broadly representative group of people that forms the deliberative mini public.

So we break that down a little bit because in some ways both are different aspects of trying to reach

Jenny Stefanotti: more people. Exactly. And why would we want to do this? What are the benefits of scaling?

Claudia Chwalisz: I think a lot of people who are interested, particularly in this, are driven by arguments around, like, for example, legitimacy.

I think this also links up to our notions of democracy as elections of wanting more people to be involved, conferring a greater sense of legitimacy on decisions that might come out of a process, even if that might be at the expense of other characteristics that might contribute or detract from legitimacy as well.

But still there's, I think, a set of arguments and thinking that if we have more people, this will be more legitimate. I think there's another set of [00:39:00] arguments around a potential benefit for collective intelligence that if we have a larger, more diverse crowd of people, that also has a benefit in terms of the quality of the kinds of ideas and things that they might come out with.

I think there's also a sense, but I think this one is probably widely shared across the five, but put forth a lot by people interested in scaling out, is that because we know that there are benefits for people of deliberating, so a lot of things that we have talked about, this is quite a transformative process.

People's sense of agency gets ignited, people are more engaged, they trust one another more, et cetera. There's a sense of, well, how do we make more people feel those benefits? And so by trying to increase the numbers basically. So those are all, I think, rather appealing arguments to a lot of people who are thinking about that.

Jenny Stefanotti: Yeah. So just to underscore that, that's. More legitimacy because more people are engaged, better decision making because more people are engaged and you have better collective intelligence. And also just the kind of benefits of directly [00:40:00] participating for trust in government, et cetera, because more people are doing that.

And here you mean both actually in the deliberative process and informing the deliberative process. What is all the information that the deliberative body is consuming to make decisions not just based on their own value set, but the way that they're representing everybody else so we can have more indirect engagement through contributing to an opinion poll about a particular topic that your represented deliberative the body is considering.

Right. Okay. So let's talk about how AI can, can support this type of scaling.

Claudia Chwalisz: Yeah, so I think, well, there's a few different ways I'd say that a lot of people who are interested in this aspect of particular have been experimenting with. And again, this is where I think our conversation earlier, distinguishing between in-person and virtual is interesting because I think I would say most if not all of the experiments have been purely virtual with a sense of using technology and also using AI to help with things like for example, instead of having a human moderator, which [00:41:00] is more expensive and you need more of them and you need to train them and so on, you could train an AI to moderate conversations instead.

So I think we've been seeing quite a bit of that, not just in these kinds of processes, but I think also when they're happening online, you tend to have recordings that are then using AI to be able to synthesize or draw out kind of sense making conclusions or other things from the conversation data that's happening.

So that might be shared with people in real time or after, or might be used for research purposes. But I think. That's not limited to the scaling out, but it tends to coincide with it because of the nature of how those kinds of approaches have been applied. And I think the other kinds of, well again, we're talking about one dimension at a time, but there's often like an intersection and, and one of the other things that we've seen with them scaling out approaches that they've also been used, maybe more connected to what we've called scaling up, which refers to [00:42:00] scaling to higher uh, levels of governance.

So also trying to think about what would deliberation at a global level look like. And so again, to try and have some greater sense of representativeness at that scale, including greater numbers. And so things like using AI to help with translation and interpretation in real time has also been another application that we have seen in those kind of contexts.

Jenny Stefanotti: Just to underscore for scaling up, obviously AI can help us synthesize across a lot of inputs for sense making for visualization just makes that process more efficient in the deliberative process. For the indirect component of it, just you can build tools to integrate voice across large swaths of the population, and then again synthesize and present that to deliberate body.

And then facilitation itself where it sounds like there's some experimentations with pure AI driven facilitation, but I'm guessing your point of view is that AI is a facilitation buddy makes more sense than AI is [00:43:00] replacing humans on the facilitation front because there's something fundamentally human about that.

Claudia Chwalisz: Yeah, it's a good extrapolation of some of my thinking. I don't think we say that explicitly in the paper. And I think the other element of it too is. Going back to what problem are we trying to solve? I also think that more people knowing how to facilitate conversations well is actually a very desirable skill we should be trying to spread in society, rather than thinking about how could we get AI to replace this?

I dunno if you were going to ask about this, but I think there's also all the trade-offs involved with these kinds of approaches where you can theoretically scale, let's say a virtual deliberative process to let's say a thousand people, for example. But within those thousand people, you're not going to get to know everybody individually.

There's various trade offs involved for the trust and the group building that happens. You can't do sensemaking in the same way. Of course, you can rely on AI to help support this, but in a process with less people, it'll be more driven first by the people who are playing a role in that. And it can, and I think should [00:44:00] be supported by AI for different parts of this process because we're also trying this in some of the smaller scale ones, but.

You have a much greater reliance on it in a larger process, and I think we also need to take into account some of the claims that are being made about more people for the same cost of having designed this process and so on. But actually, if you're still interested in making this accessible that people are paid, that you're providing the kind of support needed to people who might not have the computer at home and so on, it's definitely not a cheaper process if you wanna do things that way too.

So I just kind of wanted to make sure that was said at some point in this podcast as well, because I hear that quite a lot.

Jenny Stefanotti: Yeah, and this points to what we were teeing up, which is the fundamentally human elements of that and some of the inherent limitations to scale and what the trade-offs are when you start to use AI more when you're moving into the gray area or scaling up more when you are facing those kind of fundamental constraints to scale.

I appreciate that. And then you talked about it, scaling up is moving from [00:45:00] largely local. I mean, I think most of the citizens assemblies that you see at the local level, some at the national level,

Claudia Chwalisz: yeah, most are at the local level. I think 65% of the OECD examples of like around seven, over 700 examples are local.

There are a number of regional and national ones as well, and a small handful of ones that, well, actually at the EU level, there have been more and more transnational ones, but also actually largely in person. A number of them have been done online too, but again, for all the reasons we've been discussing, it's not always easier or cheaper to do it online.

Jenny Stefanotti: So, yeah, this specifically, when you talk about scaling up, you're looking at transnational governance, global governance.

Claudia Chwalisz: Yeah. Transnational governance. Global governance. There's been a number of people interested in ideas like a global assembly, for example, on issues that are global in nature, like climate change or actually issues related to ai.

In some ways, you could think about having global deliberations around them as well, even though I actually think some of the more local deliberations on such things and how are we actually [00:46:00] applying this in specific context make more sense than the very, yeah, large, big picture ones at a global scale.

But still, anyways, there is a community that's interested in these kind of scaling up questions. It's not the main focus of my own personal interests or what we're doing at DE next, but there are people working on this, and I think I already touched on the main tech AI applications.

Jenny Stefanotti: Yeah. Which is really an epistemic issue with translation, where AI can really help with that.

Yeah, yeah. And then scaling across.

Claudia Chwalisz: Scaling across. Yeah. Well, in our pre-chat, you rightly noted that some of these terms are not always the most intuitive. I take your point. It was hard to come up with all the names, but by scaling across what we mean is taking the beauty of the small scale deliberative process we've been talking about and wanting it to spread maybe is a better word, so spread to more issues in one locality, in one region, in one country being used by more ministries, for example, to more cities, to more countries, to more regions that are [00:47:00] using this, but also across other institutions.

So we talked about this at the very start, how do we also spread these ideas to museums and universities and schools and. Other kinds of organizations that can be applying these principles as well.

Jenny Stefanotti: Yeah, no, I really appreciate that. And this kind of gets to the point we made earlier about these are deliberative processes, they're not necessarily in public institutions, but when we talk about democratic resilience, we're particularly interested in the application in public institutions.

And so yeah, I think you stated that well, just more issues, more places. Right now it's happening more and more, but it's still quite limited relative to the full scope of public decisions that are made or decisions that are made that affect people's lives. And then, and where there then is AI interesting to you?

Claudia Chwalisz: Yeah. So I think there in some ways, AI to me is interesting in that before stage of the process from the perspective of, I think part of the challenge of scaling across comes from the fact that even if somebody wants to do an assembly, they get excited about the [00:48:00] idea. There is somewhat of a fairly steep learning curve, and the process is somewhat complicated and there are many micro steps to getting it set up and off the ground.

And at the moment, there are resources like our Dem Next assembling and Assembly guide that have tried to break this down step by step and have given people templates that you can download for different parts of this. But at the end of the day, that's still a static website with PDFs or other documents you can download.

So I think that we're at a stage and this is something we're actually quite interested in at doing ourselves at Dem next, using that as so guide as a bit of a starting point as well of how could we turn that into something that's a much easier to use accessible tool. I still think, again, I stress there's still always gonna be some relational work involved of making this happen, but I think you can use technology in really interesting ways to just make it easier and that part of the process, you could surely make it a bit more cost effective in terms of how much time is spent [00:49:00] on things that you could have much more easily replicated when they're automated process around things that you're just repeating time and time again.

Jenny Stefanotti: Yeah, I think a hundred percent, particularly when you think about this in the context of just budget constraints that you're likely to face in local governments, or particularly if you think about budget constraints that you're more likely to face in lower income countries that might be interested in adopting these models.

So scaling from the perspective of reducing friction to adoption, both in terms of process, expertise and cost itself. So. I'm excited about that one for you too, and I'm excited to hear that you're building it, so I look forward to seeing that evolve. Okay. Scaling deep. I like this one.

Claudia Chwalisz: Yeah, so scaling deep, it's really the emphasis on how do we ensure that when these assemblies or deliberative processes are happening, that they actually have impact.

And in some context, and I guess especially when we're thinking about this in a government context, how do we also institutionalize and make these more permanent, not just one-off assemblies that depend on [00:50:00] political will, because that's another way to be scaling. If this is something that has become a new permanent institution like it has in Paris and Brussels another places, then it's like, well, every year there's a new part of a hundred people who are randomly selected, and this just becomes a new part of how democracy works.

And that's probably just a starting point for me. My vision is that there's actually more assemblies every year with more people that get institutionalized and anchored to different parts of a how a government works, but. What we've seen are good starting points, but I think, yeah, so that's another vision of how do we scale by making this something that happens on a recurring basis.

Jenny Stefanotti: Well, I think that that scaling deep in terms of institutionalization is a really interesting point. Let's put it back in the context of democratic resilience. So part of this is that it's a one-off assembly. You don't set it up unless it has political will and buy-in. So there's less power decentralization than if it's a permanent body where the agenda setting is actually happening from the deliberative bodies themselves.

So the [00:51:00] citizens who are representative are setting the agenda. That's a fundamentally different distribution of power, which gets into democratic resiliency, which I think that's a really important point to make There.

Claudia Chwalisz: And I think another point related to resiliency in particular from what you were just saying though, is that you can see how a one-off assembly can have a really important impact on one policy issue being changed or affected by it.

One group of people being transformed by having been part of the process, but one assembly was 50 people in a place is not going to transform the democratic fabric of that community. It's only going to be, if this is something that actually becomes recurring, and again, across many different institutions within the same place, that you can actually start to imagine a vision of this, having that more transformative effect for democracy, for a more resilient citizenry as well.

Mm-hmm.

Jenny Stefanotti: Yeah. Again, I think in terms of legitimacy, more and more decision makings happening in these bodies, redistribution of power is just really critical for the [00:52:00] institutionalization point that we're making here. Mm-hmm. So I, I just, I really appreciate that, but also integrating more with society writ large, because you talked about stronger connections to the public sphere also within scaling deep.

Do you wanna say more about that?

Claudia Chwalisz: Yeah. Well, I think when something becomes not just this one-off tool that's being used, but actually seen as an actual democratic institution that has relationships, therefore with all the different kinds of stakeholders, which includes elected officials, the public administration, but also all the wider group of stakeholders that include businesses and civil society groups and others.

That's also part of what becomes transformative as well, that it starts to change the relationships people have with one another as well.

Jenny Stefanotti: Yeah, I appreciate that. Okay. Then the last one was scaling in. So this is about the impact of improving quality, hence impact of these deliberative processes. So let's talk about that one.

Claudia Chwalisz: Yeah, this was probably the hardest one to name. What makes the most sense? [00:53:00] Uh, so we were trying to think, well, if you're going like kind of more so into the process itself, but scaling in this also goes to the goal of, okay, it, it's not just that it matters that we have more of these processes happening or more people in them, but that these are actually higher quality and good deliberative processes that are taking place.

And so what are the ways in which we can be making changes to the design of these processes? And also I think the technology lens is particularly important for this dimension. How can we be using technology in useful and interesting ways? And we've alluded to some of them already actually in that information and learning phase, for example.

In perhaps aiding some of the synthesis work, visualizing some of the conversation data, being able to even share some of the voice conversations that have happened in kind of more creative ways with the wider public. And being able to use AI to help us trace, for example, the recommendations at the end of a process.

What was really the genesis of those ideas. And then you can have conversation highlights [00:54:00] that help us get a sense of like how people got there. And I think there's a richness to be explored. And I think, again, we've only just started as a field to be thinking about what are all the different ways in which we can be using AI and other technologies more broadly to help make these processes higher quality.

Jenny Stefanotti: Wasn't that part of that? Also just the outputs to policy and making them more effective or just re reducing frictions to adoption of policy recommendations. Wasn't that part of the in also and how AI could be supportive of that?

Claudia Chwalisz: Yeah, I think so. I mean, I think that kind of goes across both like deep and in in a sense of like, yeah, part of deep is also about how do we make sure this has more impact too.

So I think. I think that's why we really stress, these are not mutually exclusive, but there is a value in separating them out to have that more nuanced conversation.

Jenny Stefanotti: Just giving us a framework to have these more nuanced conversations. Yeah. No, no, no. I super appreciated that. Out of all of this, is there something that you're just most excited and enthusiastic about?

There's so much and I just, I, again, I really appreciate how the paper just gives us a landscape to have a more [00:55:00] nuanced and substantive discourse about trade-offs and plus and minuses, and even just mapping out all the tech and where the tech is interesting, which your paper talked about the need for, but what are you stoked about within this landscape?

Claudia Chwalisz: Well, I think what I've talked about is the fact that there is really a growing community, both amongst technologists but also around the deliberative democracy community that I think has been growing slightly. 'cause I, many of them have maybe been on a slightly more skeptical side about wanting to even use or explore ai.

But there is a growing group of people across practice, across technology who are interested in building new technologies and testing them and exploring things and having conversations together. Also about. How do we also have some form of shared principles around when and how do we deploy some of these technologies within a deliberative process as well?

Because I think as we go into this territory, there's more and more of a need for that too. So I think the fact that there's more people, I guess also at Dmex we've been trying to play some role also in collaboration with others [00:56:00] to be convening a little bit more of this growing community to have spaces to work on this.

And I think on a personal level, some of the things that I'm working on that build on this paper are, I guess, the things I'm interested in since I'm spending time doing things on them. So one goes more into this AI dimension. So in collaboration also, again with Sammy McKinney, but also with Kyle Redmond from the AI and Democracy Foundation, we're working together to, like you asked me at one point about how do we break down and map the deliberative process.

So I gave you a super high level overview because we're on a podcast, but we are. Nerds, but also wanting people to build stuff. So we're going into a very detailed, like let's break down the micro tasks at each of these stages kind of work and are going to be convening together folks from the tech and the deliberative democracy communities to actually do that collective work at figuring out, okay, what already exists or can be used, or where are things maybe duplicative in nature?

Could we collaborate and reduce those duplications? What needs to be built and hasn't been built yet? So I think that next phase of [00:57:00] what happens next is really exciting to me. And then the other bit that I'm really excited and interested in is, again, with Sammy McKinney, we're going into a deep dive into this civic infrastructure part of what's needed for scaling.

So we've spoken about this a little bit less so far, but this gives me an opportunity to bring it up actually. Yeah, it's important. 'cause I think precisely in the wider community, I would say most people are really focused on AI and how AI can. Be helping this goal and objective, and again, I share that excitement, but I also recognize that there's just a limit to what AI alone can do if we're interested in scaling.

So what we mean by civic infrastructure are these kinds of organizations that have had a catalytic role. In their regions to help deliberative processes. I would actually say in that broader sense to spread within their regions at often quite a quick pace and again, often beyond just government institutions where there's now been a kind of weaving into other kinds of organizations in those regions.

So we're doing a deep dive into Aaron [00:58:00] Zu lab, which is a democracy innovation lab in the Basque country in Spain, and an organization called, we Do Democracy in Denmark. And we're going to also look just a little bit less in depth at a number of other organizations with similar characteristics. But these are the kinds of places that, well, they have a physical space where they are doing a lot of convening, gathering, very human relational kind of work, where they're documenting and evaluating and disseminating things and a lot of outreach and connecting across the different communities of people, across politicians and civil society and practitioners and so on.

To weave a stronger foundation of people with the skills, knowledge, shared advocacy for some of these ideas. And so we wanna put a bit of a, a spotlight on this, but also to actually genuinely understand better what has made these organizations so successful in what they've been doing and what is context specific and what can be transferred or replicated in other places, places as well.

So that's the goal of that work. Yeah,

Jenny Stefanotti: it just makes me think about how do these [00:59:00] new technology span the solution space for this governance innovation, looking at technology as a component of a broader solution that includes new human processes and human elements. And again, I just really appreciate that really critical point that you make in the paper, that it's not technology alone, it's technology alongside the set of critical civic infrastructure.

And then what are our design processes that help us discern what those interventions are that are interesting? Going back to where we started with the papers, why are we here? We're not hammers looking for nails. That's how we design interventions that nobody adopts. So how do we put this in the political context and what our overarching objectives are, and understanding the human elements to design upgrades to our deliberative processes with the new solution space that AI creates for us.

Yeah,

Claudia Chwalisz: definitely. Very nice synthesis, actually. You have all the different threads there.

Jenny Stefanotti: Maybe I just have one question to close, and then I, I wanna get you off to dinner. I know you've had a long day, which is the other side of the coin with respect to [01:00:00] AI and democracy. We're talking about democratic resilience, and part of what has led to the decline in democratic institutions globally is technology and the information ecosystem and how that's been impacted by social media, in particular, disinformation and misinformation, how we're more polarized than ever before.

There's the downside of AI and democracy and democratic resilience, which is what we're here talking about, which is more disinformation, misinformation to just continue with the frames that we were just using. Expanding solutions set for bad actors to be disruptive to democratic processes. So. I know this is taking us out of the specific wheelhouse of your work in this paper, but since you're a resident expert on democracy broadly, and again your work pre getting into this was looking at populism and the rise of populism, I'm just curious about the risks of AI around democratic resilience and if you have any thoughts around ways to mitigate those risks.

Claudia Chwalisz: Yeah, for sure. And I mean this is also where I recognize some of the limits of my own in depth expertise. And also [01:01:00] just to stress. I also by no means think deliberative processes or some sort of silver bullet to all the democratic problems we face. So they're not gonna address all the things that you just talked about.

I think there's some interesting interlinkages. So that's probably the thing I'm most well placed to make some notes on. 'cause I think we're so rarely in a shared information space, but actually an assembly creates this, where this group of people is all accessing the exact same information to come to then be able to have a deliberation on that issue to together from an informed basis.

Um, so again, I think we just need more of that. And I also think that. From my familiarity with some of the literature on disinformation and misinformation, there's obviously things that need to be done to address some of the symptoms of these problems, but the deep roots of it are also interlinked with the deep roots of the problems we've been talking about.

Like people who are more prone to share disinformation are people who are lonely and feel disconnected and not part of their communities, and therefore also more likely to go into conspiracy theories and so [01:02:00] on. So again, there will always be some work to be done to reduce the root problems as well as the symptoms of, well, what do we do?

And once that information has been shared as well. And I think being able to think about actually the linkages between these different areas of work is quite interesting. And I know people are much more knowledgeable than me about some of the other kind of tactics and things that need to be done as well, but.

I have been interested in that. I guess the connection between some of those challenges too, and I'm very aware there are also quite a lot of new types of risks and with deep fakes and other things, and I think just people generally, uh, finding it hard to know what to even believe these days. Um, that's also not going to be a challenge we solve overnight or with just one intervention.

Either.

Jenny Stefanotti: Your answer got me thinking about the cultural implications of a proliferation of deliberative democracy in the way the AI can enable that because I think we're, we're in a cultural context of polarization and really deeply steeped in our own point of views and [01:03:00] confirmation bias. I mentioned this is part of a cluster of conversations and one of them is on like regenerative civic engagement and we talked about why don't people engage civically and we have a culture where there's not a lot of civic.

Discourse to begin in with. Particularly in the US we're not talking about necessarily a lot of these issues that affect us, particularly locally. So I think it's interesting to consider the ways in which scaling deliberative democracy leads to a cultural shift in just our own trust in public institutions, just in engaging and thinking about these issues and informing the decision making bodies that we actually trust, such that there's more of a cultural shift towards curiosity rather than the tech fuel drift towards polarization and filter bubbles.

Because I think that your point around a shared information ecosystems is really important here. And so I think there's an interesting cultural was like Tristan Caris gave this TED talk about AI this year. The piece of it that I thought was more interesting was this inquiry around how do we create an immune system.

Against the [01:04:00] pernicious potential effects of ai. And so I think it's interesting to think about the cultural implications of proliferation, of deliberative democracy around kind of immunity to more polarization, more disinformation, just more of a inherent civic engagement. And in curiosity, is it antidote to the polarization and bubbles that we see online, but interesting times for sure.

Claudia Chwalisz: That's where the civic infrastructure piece I was mentioning is also related to that actually. And the focus of this paper was really on deliberative mini publics. But the focus of that paper is broader. And also about how does that also help instigate a broader deliberative and democratic culture. And that's where that interest in the physical space that these organizations have and the role they play, where it's not just a conference center or anything like that.

There's like a space where it's also, again, beautifully designed and has some welcoming feeling. But there's also, at least I'm thinking, for example, the one in Copenhagen, there's also now Copenhagen's Best bakery that's there. They have a bar. There's a different coworking space. [01:05:00] There's a wider public space in the middle of that people can come and use.

And there's also been a very active effort of the people from that organization to go and knock on all the doors of people in the local community, like hyper-local community, let's say, where they're based and involve them in different initiatives they've been working on, literally pulling off concrete from the road and planting new trees together, and quite literally activating a sense of agency as citizens, as well as creating that sense of connection, the spaces for dialogue and conversations to happen.

So I'm really interested in that broader sense of things too.

Jenny Stefanotti: Yeah, and that really ties to the regenerative civic engagement conversation that will compliment this, that we're releasing around how do we create. Civic experiences that are regenerative for the people who engage in them. That means they wanna come back because they're left enriched by that actual act.

And, and so Indie talks about the simple things like, well, what if you feed people? Why don't we feed people? You wanna talk about accessibility? What about the people that can't go because they have to make dinner for their kids, so what if you [01:06:00] provide dinner for their families? And so, yeah, there's so much richness to this conversation.

I'm just so grateful that you were up to come back on. I'm just also super grateful to be plugged into what you're doing because you're at the forefront of this conversation, and I'm super excited about the subsequent papers that you'll be releasing, and I'm just thrilled to be able to provide a forum for the dentists and audience community to get the synthesis of the work that you're doing.

I highly do recommend taking a look at the paper because it's very accessible and a relatively quick read of, of course we'll link to it in the show notes, but. Yeah, just thank you so much for coming on and making the time, and I'm just, again, I'm just really thrilled about the work that you're doing, and I'm just really stoked to have a front row seat to it.

Claudia Chwalisz: Oh, super. Well, thank you, Jenny. I'm also really grateful also to be part of this denim community and the fact that you're creating the space for us to have these conversations. So thank you for having me today. And yeah, I'll be curious if there's also some other questions and engagement with other people who have listened to this after.

Jenny Stefanotti: We'll have a community conversation, so people who are listening who wanna tune into that, you [01:07:00] can reach out to me on the website or just respond to the newsletter. All right. Well, let's let you get to the end of your day. I know it's been a long day for you in Europe, but it's great to see you again.

Claudia Chwalisz: That's okay. Yeah. Thanks, Jenny. Great to see you as well. Thank you.

 

[OUTRO]

Jenny Stefanotti: Thank you so much for listening, and thanks to Scott Hanson, also known as Tyco for our musical signature. In addition to this podcast, you can find resources for each episode on our website, www.becomingdenizen.com, including transcripts and background materials for our most essential topics like universal basic income, decentralized social media, and long-term capitalism.

We also have posts summarizing our research, which make it easy for listeners to very quickly get an overview of these particularly important and foundational topics on our website. You can also sign up for our newsletter where we bring our weekly podcast to your inbox. Alongside other relevant denizen information subscribers are invited to join our podcast recordings and engage with the Denizen community in our [01:08:00] online home, the Den.

We're partnered with some incredible organizations at the forefront of the change that we talk about. We share announcements from them in our newsletter as well. Finally, this podcast is made possible by support from the Denizen community and listeners like you, Denizen's content will always be free, offering Denizen as a gift models, a relational rather than a transactional economy.

Enabling Denizen to embody the change that we talk about on this podcast through the reciprocity of listeners like you that we are able to continue producing this content. You can support us or learn more about our gif model on our website. Again, that's www.becomingdenizen.com. Thanks again for listening, and I hope you'll join us next time.