Murmuration’s Chief Marketing and Operating Officer, Michael Slaby, sat down with author Hélène Landemore, the Damon Wells ‘58 Professor of Political Science at Yale University with a courtesy appointment in the Department of Philosophy, Distinguished Research Fellow at Oxford’s Institute for Ethics in AI, and Founder of the Governing X conference series, for The Exchange, our interview series featuring influential thought leaders, organizers, advocates, and others who are shaping the future of civic life.
This interview has been edited for length and clarity.
Michael Slaby: Hélène, I’m so grateful for you being here. To start, I’d love to hear a little bit about your story—how your thinking, going all the way back to your book Democratic Reason, has helped us reimagine what healthy, vibrant, open democracy means and could look like in modern life. What has your attention at this point in your work?
Hélène Landemore: If you want to start at Democratic Reason, which I guess is the starting point intellectually, I was interested in this idea of collective intelligence as an argument for democracy as a decision rule and a regime form. That led me to think about the mechanisms through which we aggregate our individual judgments, generate new approaches and solutions, and make choices about which ones to pursue.
I began thinking that maybe we’re not optimizing for collective intelligence in the current version of democracy that we’re familiar with
Social scientific evidence suggests that groups are smarter when they are more cognitively diverse, and that this matters more than simply bringing the best and brightest together. So I began thinking that maybe we’re not optimizing for collective intelligence in the current version of democracy that we’re familiar with, which is electoral democracy. It was born in the eighteenth century, at a time when we were living under monarchical regimes, and was seen as a major improvement because it asked people for their consent and focused on elections as a way to choose leaders.
But the people designing these systems still thought that if you wanted a smart parliament, you needed to put the smart people in it. They tended to focus on the wealthier and more educated parts of the population and then choose the most competent among them. I saw a mismatch between what contemporary social science tells us about collective intelligence and the way we select the members of our key assemblies. If we wanted to maximize the collective intelligence of our parliaments, maybe we should choose their members by random selection.
That led me to Open Democracy, where I explored what an actually smart democracy might look like. Through researching the Icelandic case study for Open Democracy, I realized that selecting representatives by lot isn’t enough. You also want that group to receive feedback from the larger public, both because you might miss perspectives and information and because of questions of legitimacy, equality, and inclusion.
Then I became more involved in actual experiments, including the French Citizens’ Conventions. I wanted to tell a story that was more humane and less focused on the epistemic dimension, and more about the emotional and relational dimensions of these assemblies. That’s what led to Politics Without Politicians. Despite the title, it’s not really about politicians so much as the kind of politics you can do when they’re not in the room.
Michael Slaby: What becomes possible when we deprofessionalize representation? One thing I see emerging from the invitation to greater participation and shared power is that, even when we’re still looking at representative expression, the fluidity of access to power invites more of us to participate in the use of power—not just in who holds it.
Our dysfunction around the choice mechanism has led to the professionalization of people in representative positions, rather than inviting everyone to participate in governing, solving problems, and exercising our collective wisdom and judgment. I love the increase in agency that your work invites us to consider. How do you think about that relationship between agency and representative democracy?
Hélène Landemore: I’m glad you say that, because one of the criticisms I often receive is that lotocratic models or assemblies chosen by lot would entail a lack of agency for the people who aren’t selected. That’s why I think it’s important to make clear that I still want people to vote and have agency in campaigning around such moments of mass voting (whether they’re regularly planned top-down or spontaneous citizen-initiated referenda). I think those moments of mass voting should be about bundles of issues, programs, or one-off issues, rather than about selecting representatives.
My argument is not only empirical but also theoretical. Even if we could achieve the best version of representative democracy, it would most likely still fail to capture enough of the cognitive diversity of the larger group. That’s why I think these systems struggle to generate the policy platforms we need. In the United States, you have two platforms, roughly on the left and the right, and they fail to cover the vast range of options that people are actually asking for.
Michael Slaby: Anything that excludes any meaningful part of the populace is going to condemn itself to blind spots that ultimately matter. You’ve also written about the distinction between agenda-setting and final say. Could you explain that distinction and why it matters for how we think about power?
Hélène Landemore: Power is often assumed to be located in the final say. If voters get to choose representatives and the policy platforms associated with them, or even if they get to vote in referenda, then they have power. That’s true, but the power to elect representatives or the power to say no is still different from the power of being in the room and setting the agenda. Agenda-setting is the “second face” of power, according to a famous analysis by Peter Bachrach and Morton Baratz. And sure, you can try to do that from outside the room, through the party system as well as campaigning, canvassing, or writing letters to representatives, but we know how limited those mechanisms are. Even at the level of the ideal, it’s a limited form of democratic agenda-setting.
There’s also what some have called the “third face” of power ideology (per the intellectual contribution of another political scientist named Steven Luke)—what’s acceptable, what kind of future we can imagine, and what policy platforms are considered realistic. People are often more on the receiving end than the shaping end of that process of defining the Overton window.
The more I think about it, the more I believe an authentic democracy would put at least two out of the three “faces” of power in the hands of the public (the third one, ideology, is perhaps the one no one in particular should control). Elites would be employed and mobilized in the service of those forms of power, rather than concentrating them. Right now, agenda-setting power and even the ideological power component are largely controlled by elites, while only some of the final say is left to the people, after much of the process has already been determined.
Michael Slaby: One of the things that captivates me in your work is that, in conversations about civic renewal, there’s often a focus on electoral reform, or on bridging and bonding through social cohesion and deliberation, but those conversations can become disconnected from power.
Citizen lotteries and assemblies offer tools for participatory power. But one of the challenges is that the experience of a citizens’ assembly is not always connected to an institutional lever of power. The findings and deliberations can be separated from the final decision and implementation. In your experience, do you see places where municipalities, states, or countries are actually giving citizens’ assemblies the power to be the final approval mechanism?
Hélène Landemore: I now think the first French Citizens’ Assembly on Climate was actually trying to do just that. President Macron made a promise that whatever came out of the convention, he would take it directly to a referendum, a parliamentary vote, or direct regulation. It was not letting the assembly decide but setting the agenda in such a way that the downstream decision-makers were likely to follow suit.
He didn’t fully keep his promise, of course, but the promise was there. Unfortunately, the citizens in the assembly were given the option of a referendum, which was, in a way, the most promising route for their recommendations, and they chose not to pursue it. They were afraid that people outside the assembly wouldn’t trust them enough to support them, that the referendum would fail, and that it would doom the future of citizens’ assemblies. So they chose to trust the government and the parliamentarians and gave them their 149 recommendations.
Those recommendations were considerably diluted by ministers and parliament, so the causal effect of the assembly on the law was greatly diminished. And afterward, the government passed a law placing citizens’ conventions under the responsibility of an existing legislative chamber. So they’re no longer as autonomous or self-standing; their power has been, one might argue, domesticated.
In Connecticut, we’re running a citizens’ assembly on property tax. Unlike in France, we started with no firm commitment from anyone to implement the recommendations. But we did get a courageous commitment from the current state comptroller, Sean Scanlon, to take the recommendations and bring them to a legislative hearing. It’s a loose docking of the citizens’ assembly to the system, but it’s a form of connection. We’ll see what happens after September 26.
Michael Slaby: There’s so much talk in American political discourse about polarization and whether we have the interrelationship and deliberative skills to be productive in a shared space. In Politics Without Politicians, you talk about a good assembly being one where the shyest person is comfortable speaking. Could you talk a little bit about what you’ve learned about how these conversations work and what helps create that kind of space?
Hélène Landemore: We’re all familiar with conversations where the loudest person, the most opinionated person, or the person with greater status or rhetorical skills ends up dominating. The rest of the group is nodding silently and hiding their true opinions. As a result, the collective intelligence of the group is lacking.
The idea in citizens’ assemblies is to make everyone comfortable, especially people who are used to being silenced, overlooked, or invisibilized.
The idea in citizens’ assemblies is to make everyone comfortable, especially people who are used to being silenced, overlooked, or invisibilized. In the book, I use a phrase from G. K. Chesterton: “All real democracy is an attempt (like that of a jolly hostess) to bring the shy people out.” That captures what I’ve seen in these assemblies.
There are all kinds of facilitation techniques to bring out natural introverts, but also people who are made structurally shy by their life experiences. You can change the format, break people into smaller groups, or use chat instead of speaking on camera. The norms are also explicit, and people can appeal to them. People are told to listen carefully and let other people speak, so when someone talks over others, they’re the ones being called out for abusing that privilege.
These processes are far from perfect and still reflect inequalities in the larger society. But of all the settings I’m aware of, I think citizens’ assemblies are among the most welcoming and inclusive.
Michael Slaby: One of the things we see in our research is distrust in institutions and people struggling with their relationship to civic life. We’re living through a moment of intense populism. But beyond expanding economic inequality, we’re seeing expanding inequities of political power. Many people feel that we’re pretending to share power without actually inviting people to express it.
When people have the opportunity to participate in spaces like citizens’ assemblies—to be included in the expression of shared power and work together, even when the process is imperfect, or the recommendations aren’t fully implemented—does that experience change their relationship to community, institutions, and civic life?
Hélène Landemore: I think it does a lot. It can help depolarize people, bring them back into the fold of regular electoral democracy, and even increase trust in politicians. Although I tend to present citizens’ assemblies as an alternative to politics with politicians, they can actually help re-legitimize politics with politicians.
We’ve seen this in Ireland, where citizens’ assemblies became a routine part of the lawmaking process. The popularity of politicians increased, and Ireland became the European country where people were least likely to say they wanted to replace elected politicians with randomly selected people. The same thing happened in France after Macron organized a national debate and citizens’ convention. His popularity increased again, at least for a while. Was it specifically because of the randomly selected assemblies, or because people felt that he was making a good-faith effort to listen? I think people are willing to give a lot of goodwill when they feel they’re being heard.
At the individual level, I talk in my book about Jules, who was deeply disaffected with politics. When he arrived at the French Climate Convention, he wouldn’t even leave his carry-on at the door because he thought it was a scam and planned to leave at any moment. By the end of the nine-month process, he had embraced the community so much that he returned home and ran for election.
His self-esteem and faith in the collective had been transformed. He still considered himself a climate skeptic, but he voted for the recommendations that came out of the convention. He said he felt loved, made friends, and found a renewed sense of purpose, which he redirected toward, or at least, inside the existing political system.
Michael Slaby: It’s striking how the experience of love and belonging can transform our relationship to trust and institutions, and disconnect those things from the routine experience of politics and civic life.
You also mentioned that both populist and elitist approaches tend to promise outcomes to improve people’s lives without necessarily giving them a voice in that process. How do you think about that distinction?
People certainly want outcomes, but they also want a voice. In fact, I think you don’t get the right outcomes unless you give people a voice as well.
Hélène Landemore: What frustrates me about both the populist and elitist answers is that they focus on the “for the people” component of democracy. They promise solutions for our welfare. The tech oligarchs might say, “We’ll give you universal basic income,” while they continue to collect your data and make money from it. Populists might promise to lower taxes or punish the rich. People certainly want outcomes, but they also want a voice. In fact, I think you don’t get the right outcomes unless you give people a voice as well. That’s what is often missing from these approaches.
Michael Slaby: I’d love to pivot to another dimension of your work: the relationship between new forms of artificial intelligence and democracy. At Murmuration, we’re about to publish a report on how AI affects representation, power, and civic life. Where do you see both the challenges and opportunities in this current collision between AI and democracy?
Hélène Landemore: I’m an optimist. I think AI is amazing, and if used and governed well, it should serve human purposes and augment our democracies rather than undermine them. The question is how we get to a place where it’s governed well. I don’t know.
In terms of civic technology, we started using some tools during our citizens’ assembly on property tax in Connecticut. Civic tech is still in its infancy, and there’s nowhere close to the level of investment that we see in commercially oriented AI tools. But we’ve partnered with MIT and other organizations to try to connect the mini-public—the deliberative assembly—to the maxi-public, or larger public.
That’s the big challenge in deliberative democracy: how do you connect the smaller assembly to the larger public? How do you get the larger public interested? You need local media, which has largely disappeared, to report on these processes. And it’s difficult to make people care if they don’t have agency. You need to give them a purpose, such as commenting on a platform. When you do that, people actually contribute substantive and interesting feedback.
But then, on the receiving end, the assembly has too much information to process. That’s where AI tools can help with summarization and clustering, while still preserving the substance of what people are saying.
Michael Slaby: At Murmuration, we think a lot about augmentation over replacement. What are the ways these tools can make possible something that has always been enriching, but has been difficult to do at scale—like listening and processing information without flattening the signal?
Hélène Landemore: We used AI inside the assembly as well. We recorded all the deliberations, turned them into anonymized transcripts, and fed those transcripts into a tool called Voice Trace. It generated a report where every statement could be traced to particular voices, although the voices weren’t individually identified.
It gave the group a record of the discussion. It was more of an archival tool, but it meant that we didn’t flatten the group’s output into a few general statements. We preserved the nuances and could trace them back to individual contributions.
We’re still trying to figure out how to use these tools most productively, and there are a lot of methodological and ethical questions. Our participants were suspicious of the use of AI, so we told them their data would not be used to train commercial models and would remain on servers at Yale University.
We couldn’t promise data sovereignty or data ownership because we don’t yet have the legal tools to make that happen at our level of organizing an assembly. But those are questions we need to think about seriously if we’re going to use AI in civic environments.
Michael Slaby: One of the questions that has come up in our conversations about AI is what’s missing from the corpus—the human knowledge, history, and culture that may not be represented in the data underlying commercial frontier models. What’s missing, and what is the cost of what’s missing, is extremely difficult to understand right now.
But it connects directly to where we started, around cognitive diversity and the risks of exclusion. What blind spots are we building on top of, and how do we mitigate them? That doesn’t mean we shouldn’t use these tools, but it does mean we need to be clear about what requires proactive inclusion and what the cost might be of missing things that aren’t in the underlying corpus.
Hélène Landemore: Listening to you, I can’t help but think about what just happened with OpenAI scooping up the work of mathematicians by brute-forcing the resolution of a mathematical problem. They spent twenty million dollars on tokens. I was thinking, “My God, give me twenty million. What else could we do with those twenty million tokens?”
I’m sure it’s important for science, but human mathematicians were about to get there anyway. It feels like a question of prioritization. Who decides what we use these resources for? I don’t claim to know the answer, but what I know is that it cannot be just OpenAI making those decisions.
Michael Slaby: This has been a beautiful conversation, and I’m so grateful we finally got to have it. One question I always ask our new friends: Given the possibilities and challenges of this moment, what’s giving you hope for the future right now?
I think very smart people are sufficiently worried and still virtuous enough to place the common good over their self-interest, and I think they will help.
Hélène Landemore: I think very smart people are sufficiently worried and still virtuous enough to place the common good over their self-interest, and I think they will help. When somebody starts something ambitious, I’ve seen people rally around it and offer their support. Maybe it’s not enough, and maybe it’s still too soon, but I think that if the stakes are high enough, people will respond. I have faith in human nature somehow.
About Hélène Landemore
Hélène Landemore is the Damon Wells ‘58 Professor of Political Science at Yale University, with a courtesy appointment in the Department of Philosophy, and a Distinguished Research Fellow at Oxford’s Institute for Ethics in AI. Her work spans democratic theory and collective intelligence, with a growing focus on AI governance. She is interested in normative questions such as what makes collective decisions both legitimate and reliable, what forms of non-electoral representation can claim democratic authority, the place of experts in a democracy, and what democratic oversight of artificial intelligence should look like when its effects cross borders and outpace existing institutions. She is the author of Democratic Reason: Politics, Collective Intelligence, and the Rule of the Many (2013), Open Democracy: Reinventing Popular Rule for the Twenty-First Century (2020), and, most recently, Politics Without Politicians: The Case for Citizen Rule (2026). Beyond her scholarship, she has advised and designed citizens’ assemblies, including as Director of Design of the 2026 Connecticut Citizens’ Assembly on Property Taxes.
About Michael Slaby
Michael Slaby is a leader in how values, systems, strategy, and technology drive movements and organizations. At Murmuration, he leads marketing, fundraising, network engagement, and culture. Before joining Murmuration, he was a senior strategist and head of community at Harmony Labs, where he worked on accelerating media reform and transformation. He founded and was head of mission of Timshel, a social impact technology company, and was a fellow at the Shorenstein Center at the Harvard Kennedy School of Government. Michael helped lead the Obama for America campaign as chief integration and innovation officer in 2012, where he oversaw all technology and analytics, and as deputy digital director and chief technology officer in 2008.
Murmuration is a nonprofit working to transform America into a nation where everyone can thrive. We organize a network of community-focused partners and equip them with the insights, tools, and services they need to help communities build and activate power more effectively. To learn more, visit murmuration.org.
i really appreciate helene's work, loved Politics without Politicians, and great that she's getting more exposure. i am a little disappointed there wasn't much discussion about the fundamental hope of sortition: empowering a representative group that will seek their own interests. i think we sortition advocates fret and frown and worry so much about how exactly to process design and facilitate an assembly, but at the bottom of it all, if our belief is that a representative group can seek its own interest, then if it's in their interest to use AI tools, to use certain factilitators or processes, then they'll figure it out! i know the CT assembly passed on some kind of governance authority to a subset of the assembly, which is very cool. so i do think helene is aligned with this view. and by all means, i'm not opposed to developing software, tools, and processes, but their adoption should be at the whim of the sortition body. of course, we're in this kind of...introductory/experimentation phase, so what do i know? i am, after all, but one voice :)
another kinda nitpicky point, on a meta level, i do think sortition advocates are mostly seeking their own interest as a member of humanity who wants fair governance and considers all human voices sacred (or something like this). i'm not so ready to grant that we need to be like...thankful that smart peopel are willing to be altruistic...i'm more like...humanity is beginning to realize that elections aren't all that they promised to be. and, in some sense, it was kinda reasonable to give elections a few hundred years of testing before we start considering something beyond.
final complaint! i also feel like basing arguments for sortition (and civic assemblies) around something like "intelligence" can be an unsure footing. who defines "intelligence" over time? if it turns out that sortition isn't making the choices that powerful people consider "intelligent", what then? i do think helene is aware of this concern too. and in her space, and probably in this space, most people here probably have high trust of academic institutions, studies, and credentialed experts. it seems to me, at least in different parts of the USA, when people refer to "smart people", there's a lot variety of assumptions made about who those people and what they're good at, and what their goals are.
i'm sorry this is all coming off kinda negative! i'm just one voice! and sortition has given me such hope. thanks for publishing this interview.