Ovadya Metagov
Metagovernance Seminar Archive | 2025-10-21 | Unknown
Speaker 1: Oh, one second. Okay. Great to, great to see you all also. And today I'll be talking about, what I'm calling platform democracy and with the intention of how do we go how do we govern platforms in ways that don't have this problem of being run by corporate CEOs or maybe by partisan or even autocratic motivation and context for this exploration, and the resulting, like, real world...
Top Keywords
- assembly 0.009
- platform 0.008
- assemblies 0.007
- platforms 0.007
- processes 0.006
- deliberative 0.005
- representative 0.005
- representative deliberative 0.005
- process 0.004
- even 0.004
- democracy 0.004
- platform democracy 0.004
Transcript
Speaker 1
0:00 – 0:00
Oh, one second. Okay. Great to, great to see you all also. And today I'll be talking about, what I'm calling platform democracy and with the intention of how do we go how do we govern platforms in ways that don't have this problem of being run by corporate CEOs or maybe by partisan or even autocratic motivation and context for this exploration, and the resulting, like, real world work. I'll give an overview of citizen assemblies, which is the primary model I'm gonna be discussing, talk about the applicability to platforms, and then we're gonna have some q and a. And most of the things I talk about, at least my part of the work, is linked to from, Aviv. Me. So you don't need to get this whole slide now, this, this is just a preview of you'll see this again at the end, but but the goal of like the underlying question here is to answer this question of who who decides? Who should decide? And if you one second. My keyboard has stopped working. Fascinating. And my mouse. Okay, great. And if we zoom out one level, why do we care about who decides? Because for lots of reasons, but the the core a core piece here is you have these large platforms, things like Facebook, YouTube, TikTok, if you wanna call it meta, alphabet, etcetera, ByteDance, and they provide this sort of reward structure for society. But who decides what that reward structure should be, what that word structure is? That's what the question here is. And, again, zooming out one level higher, these systems are what you can call social technologies. They you know, social media, search engines, messaging, even voting systems, governance systems themselves are are ways of of of interacting with, organizational systems, organizational technologies, and also, you know, digital technologies. And they there are things that collect and aggregate and route communication, information, and attention. And we need to figure out how to structure these in order to support maybe some set of underlying goals. And going assuming at one level higher, one set of underlying goals you might have and sort of something that guides my work is that social technologies should facilitate understanding, trust, and wise decision making. Another way to think about some of this is pluralism or plurality. And when you're thinking about wise decisions, well, there's another way to approach that, just looking at what are the properties that you want in your decision making systems. And so this is sort of the framework that I've been using to evaluate potential approaches to improving our decision making systems, and especially things that app apply to, on for platforms, but it also works more generally. I don't wanna get into this too much. This is still a bit nascent, but there's also this motivation around what I see as this conflation of platform power and platform impact. And we can talk about that more later, but for now, let's dive into what we're even talking about here. So right now, we have what could be called platform autocracies, and the question is how do we move to democracy? Or at least, how do we get to a world where we're we're we're not at the whims of either platform CEOs or regulators? So let's take a step back here before we talk more about platforms. So I want you all, like, to look at and appreciate this image. I know, you know, at least one attendee here is very familiar with it, and what happened here. This is the European Parliament Hemicycle. It's the room normally used for the largest parliamentary assembly elected by direct universal suffrage in the world. But those seats the people in those seats are not members of parliament. They're ordinary citizens selected by democratic lottery, by a sortition, which we'll talk about a little bit more later, and they're paid, they've gone through an intensive process of education, they're developing recommendations that are informing or have that are that are now, you know, feeding into a process that will determine the future of Europe. There are over 20 languages represented, and this sounds sort of crazy if you haven't been paying attention to this whole movement. But it's real, and these processes are now being used around the world, impacting policies ranging from nuclear power policy to abortion, climate change across a variety of nations and at every level of of governance. So that brings us to sort of the working paper that I put out around platform democracy and sort of how how do we take how do we take this problem of social media and and social technology power and and sort of connect it into this idea of, okay, maybe there's this way of doing democracy at at very large scales. And so, like, the specific problem, like, that platforms are facing is that there's a lot of, like, these prosocial improvements that are blocked, and they're blocked for a variety of reasons. Obviously, there's there's, you know, the profit motivation that could get in the way. There's, you know, there's things like that, but there's also, even aside from the sort of self interest, there are these other blockers. And those the one primary blocker is stakeholders that have very differing opinions on what they want. And, you know, the stakeholders can themselves be partisan politicians who want to entrench their power. There's also this question of, like, well, why who are we to make these decisions for for like, to make big decisions? Like, for example, I'm advocating for recommend recommender systems and ranking systems that that bridge divides, right, that support pluralism explicitly. Who who is the platform to make that decision on behalf of their of of their their users? Well, that's a good question. They need some way of actually understanding what those what people want. And maybe even I would advocate for giving them some level of agency in order to give them legitimacy around that. And this is actually really important because this helps address all those stakeholders. If there is some deep some deep legitimacy around the decisions they're making, that that is incredibly helpful, both because that that provides a, a sort of mechanism for them to to add to it provides a way for the people who are being impacted to to actually influence the the platform itself and and the the decisions that it makes. And so yeah. Basically, if you're an ordinary person, you don't necessarily want your either your CEO or your or your politicians to to have power. And so then the question is, well, that's great and all, but what does it what would it actually mean like to have governance of platforms of for and by the people, especially when you have, let's say, a 100 platforms or or 10, and you're not gonna have elections for each one. It would like, we already know that elections we have people that find that attention, and there's a lot of apathy. There's elections are dominated by extremes. So if you have all these different elections, that would be a little crazy. And the same thing for referenda, and we've actually seen this happen. Right? Facebook has tried to run elections early or, like, a referenda early in this time, and it didn't really work. And so again, I guess the question is how do what do we do like, like, how do we take this from idealistic rambling into something that's actually concrete? And so what does that look like? It's just, you know, new democratic paradigm. The thing that I sort of was alluding to in this future of Europe photo, many of you already heard about this before. I wanted to sort of go through this briefly for those who haven't. The basic idea behind representative deliberative democracy or what's called sortition. Statition really just refers to sort of that that step around selection of the people, but you can use it for the overall process. Basically, you have some issue that's selected. Let's say, what do we do about political ads? Or how should how should how should we determine what is rewarded? Or, sorry, what should be rewarded by our by our ranking systems, by our recommendation systems? You invite people to attend in a representative way. And so what that means is it's not whoever wants to show up who who is who attends or who is part of this process. You actually there are a set of processes, and I don't wanna get into the details here, but there's some some great, great material on this, For actually getting at representative stratified random sample of of of a population. This is similar to think about pulling. It's the same basic idea. You can just do it a bit more rigorously. And, again, you're you're sort of stratifying by demographics, maybe by political ideologies, by income, all these different pieces that help you ensure that you're getting a representative population. And then you take these people, you put them in a in a room, either virtual or not, and you go through these very, these facilitated deliberation processes, where as much as possible, the facilitators there's there's a set of processes that are being developed, so the facilitators don't have their their sort of hands on on the scale here. It really is being led by the participants, and the facilitators are just providing structure for that deliberation. And then the output of that might be is a policy recommendation or decision. And so you can think of it as as a way of creating a sort of legislature on demand for a given issue. And, okay, that you know, again, that might sound crazy for those who aren't familiar with this before, but this has been done now all around the world at all scales. Here's a set of issues, for example. And I think earlier here, we had yeah. These are, like, different countries where this has been done. And, again, just to sort of hone in on the definitions and the terminology, it's a little bit confusion confusing. Sortition is the idea of giving everyone from the population an equal chance of being selected for participation. So you might only have a 100 or 500. So say if it's a, you know, global or UI assembly. Deliberation is that process where you weigh evidence, deliberate through facilitated group discussions, and you're getting input from stakeholders, from experts in order to do that. You're able to reach out to them to get information about the decision you're making if you are one of the assembly members who are part of that group that's been selected. The rep a representative deliberative body is sort of the actual body that does this, and a citizen's assembly is sort of a colloquial name for that. So building on all of that, platform democracy, again, just refers to this abstract ideal. Right? Just like democracy, government of the people, by the people, for the people within the context of digital platform. But a platform assembly is a specific approach to platform democracy that utilizes this new democratic technology, the, you know, new issues. Arguably, this was done, you know, two thousand years ago. And then I also started talking about the platform deliberative vote as a slight distinction between these. So the the only difference here is that in a platform assembly, the participants are actively developing the proposals. They're writing them themselves. And you might think this might be hard or might not work or, like, it would be low quality with random people from a population. In practice, I mean, I've seen it. I've seen sort of end to end processes, like, this really does does it really can work again, and it really depends, and requires a sort of standard of facilitation in order to to to support that. And that's its own sort of set of challenges, but it's totally achievable. But the platform deliberative vote is simply where the proposals are already predefined, and they're being sort of voted or ranked by the participants. So it's a little it's a lot less agency that's being given to the assembly. And putting that all together, these are sort of the the the sort of frames that I'm trying to bring to this versus what's already existing. And I also want to allude to the complementary set of solutions here to this problem, which are things like pull list, things like psi. These are incredibly valuable. They aren't as much end to end solutions yet for actually getting two proposals. And so I see there being a a body of work around giving them and then we'll I'll tease talk about this a little bit later, sort of getting to the place where they they can at least right now, they very much support these processes, but they aren't replacements for them, because there's a lot of sort of benefit of of having these much more in-depth, you know, forty, fifty, sixty hour participant hours per participant to really get into the weeds to actually come up with solid decisions. Right? So you can think of platform democracy as this umbrella, and then these representative deliberative bodies are one approach, which you can have an assembly, you can have deliberate and vote, and these additional digital digital governance tools that can feed into that. I can't give concrete examples on the record of how this is happening within platforms, but I can say that there are pilots that are, you know, happening now, and I think it's incredibly exciting time. And I can sort of give an example of sort of a proposal that, you know, I'm circulating with with, platform that that was investing in this just to sort of, like, okay. What does this actually look like? And what does that what is what does it look like to sign the contract to do this thing? And usually, the way this works is you have a third party convening organization, you know, and there's this democracy RD network that is sort of this nexus of these facilitators, of these these deliberative processes, also known as mini publics. It's a it's a microcosm of a public that is then going through, deliberating on coming up with sort of what what can support that that public. And so the outcomes here, you know, platform decisions being made democratically, taking into account informed perspectives, lived experiences. And for platforms, this is really handy because they're just saying, okay. Here's a policy hot potato. We don't wanna touch this. Anything we do, anyone's gonna gonna annoy us. Let's give this to the people, and they can figure it out. It's also really handy because you can go to, your powerful stakeholder, maybe your your angry senator, and say, well, you know, it's nice that you don't like this. But they actually use a deep like, a deeply democratic process to come to this decision, and this is what people came to after they're they're they're considered judgment given all the information. So there's many, many models for how to actually implement this in practice. You can think of this national, regional, global. You can have platforms running the assemblies. You can have, you know, regular assemblies on a rotating set of issues. You can have assemblies that decide what issues should be covered by platforms. This is similar to something that's happening in in Belgium and on other places where there's an agenda setting assembly body, that then spawns other assemblies. You can have legislation that requires that they run assemblies. You can have this be an antitrust remedy. You can have, a platform regulator that has assemblies as an oversight body. But I think the the takeaway here, though, is that we have this false dichotomy. This is from a survey that the Knight Foundation put out, and it's a great survey. I love that they let you sort of, actually go through it yourself and see how you compare to others. But I think it and you highlight. It's the social media companies making their policies government regulated on the other, and they're you know, maybe we can there's a bit more here. And so just to sort of, you know, what what are some interesting directions and things I'm exploring? Well, okay. Citizen assemblies, they scale to arbitrary large populations, right? Because you just take you're taking a fixed size representative subset of the population, but the number of participants stays low. What are the ways how can you dramatically scale the number of participants where it's appropriate? This isn't gonna be appropriate in all situations. But but the key is how do you get those delivered benefits? And then here's something that I'm not gonna go into here today, but their advances in artificial intelligence, particular language models, and also some things in recommender systems are very interesting in terms of of supporting that. And I think other people in the in the medical community, are also exploring that, and that's very exciting. So as a final note, this is happening. I'd love to accelerate this. If you or, someone you know is actually excited about making this happen within a platform, the more allies and more advocates and more internal sponsors, the better. You know, leadership in a number of platforms are excited about this now. I'd love to expand that and make it as easy as possible to have this happen more and more and with more and more agency for the assemblies. Might be setting up a working group to further refine this. So just let let me know if you're interested. And it's, Aviv dot me. This is my email, Aviv dot me. And I'm I'm starting to consult for for folks who are interested in helping accelerate this process, particular funders. And so and then I guess, like, or the broader sort of agenda around social technologies, and so really interested in helping support that happening. And so, yeah, let's let's make more things like this happen and bring back to that that first slide. Let's have some q and a.
Speaker 2
0:15 – 0:15
Awesome. Thank you, Aviv. Fantastic start. We've already got a lot of questions and discussion starting in the chat. Please just if you have something that you wanna raise, pop it in the in the chat, even just something as simple as I have a question, or you can raise your hand on with the Zoom raise hand thing or briefly summarize your question in the chat, and I'll I'll, bring folks in with a kind of modified, stack, approach. The first, voice we had well, first was Antoine, calling out friends in the in the picture there. That's cool. But we have a bit of a conversation here between z and Antoine here, and I wonder if maybe we could hear from from both of you. Maybe first, Zee, and then Antoine, and then and then Aviv could respond.
Speaker 3
0:30 – 0:30
Oh, yeah. I was just sort of, remarking on the fact that the lottery systems ultimately still even if they select uniformly are gonna result in potential nonuniform behaviors in terms of accepting those things and the extent to which there were programs in place. It sounds like there are to offset costs, but then also whether there's any sort of, like, empirical studies rep like looking at whether the the actual acceptance rates have non trivial variations across the demographics.
Speaker 1
0:45 – 0:45
Yeah. This is a really good question. And I think, like, one of the things to really think about with this is always to compare to the status quo. Right? So comparing to the status, like so I am not gonna say that representative deliberative bodies in their current form are perfect. I am gonna say that they have a number of really interesting properties that that make them better than the status quo along those properties, and maybe even along even, like, a a broad set of things that you might wanna have. And so compared to, let's say, voting, there's a lot of inequities in a in in voting in referenda, and especially in in in enabling people to to vote as they would like like to vote, which is, many people would prefer to have significant time to actually explore and think about who they're actually supporting and and what they would like those outcomes to be. And that is something that is extremely inequitable in terms of access to that. Not to mention access to the voting the voting infrastructure itself, time to do so and all that. So, like, that's the status quo. And then when you talk about referendums, it's even harder because there's a lot of information to process around how how to navigate, like, what are the real impacts of this particular policy. And you have a whole set of institutions around that that don't that again, like, actually decrease maybe the equity considerations around that. But going back to the core question is, like, what are the like, what do people do to actually address this within these representative deliberated bodies? It's across the board. The best versions of these processes provide, like, both, resources, like funding for their your time, funding for elder care, funding for childcare, travel, their, you know, for for the, virtual processes. So, like, let's say, the global climate assembly. They're, you know, they like, you provide the actual hardware to do that, the connectivity to enable you to participate, the training to use those systems that allow you to participate. Right? Like, they like, you just there's so many levels of of of equity that you can provide. And one of the things that's really interesting about these processes is that because you're taking a smaller subset of the population, you can invest really, really deeply in every member of the dis of the of the body, and which isn't like so you're not doing something for 8,000,000,000 people or or a 100,000,000 people. You're doing things for a 100 people or a thousand people. And so the amount of support that you're able to provide to address those equity concerns goes that much higher. All that being said, yes. Self selection is still potential issue. There isn't, one of the best reviews of sort of the overall empirics, like the the sort of research on these processes is, article in science magazine, I think, in 2019. It's referenced in, in in the paper. So if you go to platformdemocracy.com, that redirects you to the paper, which has a link to the science piece. And and that I I think that may cover that, but I'm not entirely sure. That definitely covers a lot of other the common common critiques.
Speaker 2
1:00 – 1:00
Go ahead, Antoine. Do it.
Speaker 4
1:15 – 1:15
Yeah. Maybe to to complete. So thank you very much, Aviv, for the presentation and and the discussion. I think for me, one interesting discussion is exactly the representative part. And because in itself, we know from the law of big numbers that with sortation and quote quotas, you can reach a pretty fair representation of any population with 600 to 800 numbers. But that's a problem too in terms of quality of deliberation. So many processes go for lower numbers. So there is a but at the same time, for me, saying that it's statistically representative is a problem because this is what people then expect or think. And they say, okay, but no, because you can't statistically represent all the diversity. So what we say is a waste that what you represent is the diversity of the population. And for two reasons, first, because you have the people, but, secondly, because you integrate that in the process. But every part of the process also where you actively look for opinions and views that are not in the group, And that's another way of bringing in the diversity. And maybe is the on your on your question on on on code and how many people answer? So the in the research over the the past forty years, it goes from 5% to 40% which is better than opinion polls because opinion polls now are at 2%. And it's almost now better as election because the last French election was 40%. But I think the the key here is so it's for what are the incentives and research has shown that people mainly come for political reason. They want to give their voice on a political topic when you look at their motivation. The second is the social one. So meet people, talk with people, have good time. And And the third one are the incentives, which are financial or nonfinancial. And these are the the key three drivers for that. But in in the meantime, you can reach pretty good recruitment. What is missing is the infrastructure. For example, at European level now for the conference, we had to go country by country and have a different process almost in every country because there is no infrastructure for that. So one key question today is, can we have good infrastructure for election, what is the structure for random selection sortation? And I wanted and I think the key point is, Avid, what you say is what it is, what we compare it to, because very often we are harsh with that kind of innovations because we don't take the same metrics as the thing they should replace. And so we are over criticizing those new things because we have higher metrics for those. And I think that's a very interesting discussion too, is whether you compare it to with the parliament, which is absolutely not diverse and that kind of thing. So maybe a couple of comments.
Speaker 3
1:30 – 1:30
Yeah, I guess just going back to the fact that I of introduced something that was numerical in nature. It wasn't meant to say that, you know, these are imperfect. Therefore, they're they should be good replacements. It's more that, by their very nature, you could be able to compute certain lifts. So if you have a baseline, you could look at improvements and you can also look at relative improvements. You know, in the case where you talked about it was discussed the, the sort of law of large numbers and the sort of representativeness. You know, the questions that emerge are like the direct trade offs between sort of the representativeness of having, you know, 600 people versus, like, as you factor down and you increase the quality of the deliberation, what do you lose? And so, you know, again, it's it's less of a, like, this is good or bad line in the sand and more of a, like, what are the the tools that we can use to explore the design space and, like, kinda iterate, across multiple again, inherent trade offs. Right? These are not any these aren't the class of problems that have final solutions, which means by definition, you need, like, a way of doing direct comparisons and or assessing trade off spaces.
Speaker 2
1:45 – 1:45
Maybe I'll bring in Seth who whose questions and has also been engaging in this discussion, but but whose question also raises dynamics around you know, the affordances of the assembly model. So go ahead, Seth.
Speaker 5
2:00 – 2:00
Yeah. I I guess I'd love to be wrong, but my, like, my ideology is that there's gotta be something about citizen assemblies that is worse than the status quo because they're
Speaker 1
2:15 – 2:15
Uh-oh. Are you all you all lost him?
Speaker 5
2:30 – 2:30
I was wondering if you could challenge yourself to to come up with it.
Speaker 2
2:45 – 2:45
We lost you in the middle after that bombshell of something worse that, about citizen assemblies. Can you can you share what you said after that?
Speaker 5
3:00 – 3:00
Yeah. Could could you could you challenge yourself a view to come up with what if there is, like, something that would be worse about Yeah. This and assemblies, what you think that would be?
Speaker 1
3:15 – 3:15
Well, so I think that that there's still, like, a sort of new approach in some sense in terms of we don't have the standards evaluation criteria, and they're and they're more complicated in some ways and at least, like, raw voting. Right? Like, I think it's if you compare citizen assembly process to a constitution, they're like, okay. There are reasonable levels of complexity. But if you compare it to, like, voting on an issue, it's, like, orders of magnitude more public like, there's more hyperparameters in in the machine learning sense. And those can actually those parameters can, so they just I guess, out of the machine learning sense, there's more there's more, like, things that you can that can be different across different different there's more, like, things you can toggle to to change the way that the the process works. And and and there's it's very easy to have different levels of quality of the process and and, like, poor quality processes can make it much harder to get legitimacy for the higher quality processes. Whereas elections is just like, do you know where all the votes are? Are they all counted? Is there good access to the polls? Like, at least we have different, like, different sort of ways to evaluate these things that are maybe simpler to to understand. And so I see one of the biggest challenges around this is creating a baseline of quality around particular types of representative deliberative bodies that you can that you can then evaluate. And having, like, the infrastructure for that, like, that to me is is hard. And then, I mean, the other part here is that there is real benefit to having long term engagement with the same, like, with the same issues and building relationships and, like Yeah. No. No. I'm not sure. Like like, all of that all of that stuff. Like, there is some stuff there that is important. And I think, being able to really, like, clarify what that is exactly and then figuring out how to translate those good parts into these other systems is is incredibly important. All that being said, when you're talking about a specific issue, like, what values should we have, that guide our society, guide our platform, or how do we wanna trade off? Like like like, what should like, what what key pieces of infrastructure do we need to have reasonable water quality given the perspectives of all the experts? Like, you know, all these sorts of issues. Like, those are things where you can take them to an assembly, and they provide, like, a process that has, like, a near clean input and output that is incredibly valuable in avoiding so many of the the challenges of, like, an an ordinary either political environment or a corporate environment.
Speaker 5
3:30 – 3:30
Thank you for that. If I could ask one more question, I also want to I wanna question a little bit the assumption of scalability. I I want it it seems intuitively true, but I just wanna, like, kick it a little bit to see how true it is. I'm thinking of situations where there's a very small number of people who are very expert, whose actions are a really big leverage point. You know, good ones kinda sustainable herders, like buffalo herders. You know, they're they're they're great for large amounts of land. And the chances of one of them being randomly picked, and therefore, their voice being represented in an important policy is just super tiny. And And you really would have to randomly select a large number of people to get one. And so just as a society gets larger and larger, I you know, what if we have, like, an exponential growth in the number of interest groups? Like, does that undermine, the scalability of sortition at a national scale?
Speaker 1
3:45 – 3:45
So so again. And I I would say that if it was just the assembly members on their own, yes. 100%. That would undermine it. But it isn't. Right? I see these processes as sort of there's an input which is expert perspectives, interest groups, prospective stakeholders, right, that gets filtered through this democratic, like, deliberative goo. And then out comes the the sort of interpolation of, okay. Here's what all those experts said. Here's what our experiences are of of being, like, rep like, a representative population of, like, actually being impacted by those things. So we're impacted by those buffalo herders. Right? That affects our our lives indirectly, maybe through their management of the land or through our capacity. Like, I don't know exactly about buffalo herding, but, like, I can I can think about this for for other examples? It, Or it it maybe would affect us directly, but it may be something where we because we have, you know, empathy toward the the plight of others and because of, like and and our mandate is to sort of support the this the the broader population that we as assembly members, you know, the that that you're you're wanting you're wanting and trying to incorporate that, and that's sort of what happens in these processes. And so, the the stakeholders and in this case the buffalo herders be like a stakeholder group so if they're sufficiently important that they're sufficiently impactful on this decision like the the rule of thumb is like if someone would would be impacted significantly by a decision they should be part of the sort of stakeholder process that sort of feeds into the assembly. But but so so that could the stakeholder process or the the stakeholders' perspectives, again, are filtered through that democratic, everyday person, you know, representative lens, and then outcomes decision that that that seeks to incorporate all of that.
Speaker 2
4:00 – 4:00
Okay. Wanna bring b in? Go ahead. I know Bee Bee said that they had some, audio issues, so we might have a problem here. Bee, are you there? Oh, I'm sorry. Let us know if if you're able to work it out, and we'll we'll bring you back in. In the meantime Sorry?
Speaker 5
4:15 – 4:15
Oh, we we need just, like
Speaker 2
4:30 – 4:30
Okay. We'll we'll try again when when they when they're able to reconnect. In the meantime, I I had a question too, which is just about the relationship to the, economic, design and and ownership structure of these platforms. I mean, the when you're doing deliberative democracy or or assemblies in the context of a of a presumably democratic government, there's already this framework that the the government is, in a sense, owned by the citizens. And so the the idea that you would have citizen voice having some kind of some kind of role in and and leverage in shaping that government is kind of built in all the way down. In the context of platforms, they're owned by shareholders generally. They're they have management structures already. So one could see, you know, a a kind of disconnect between creating between that power structure and these deliberative spaces, particularly if they're going to have any any meaningful power. And, you know, Facebook ran its early experiment with, like, you know, having users vote and clearly did not actually want that vote to mean anything. And you can see, you know, all sorts of reasons why it could create a kind of internal conflict and even potentially legal issues if companies are following the lead of users and not the governance of their of their shareholders. So I'm just curious about how you see this kind of model fitting into a context where there already is a power structure, and it's actually a power structure that appears to be at odds with with with the kind of governance you're presenting. And and and in particular, you know, when you're talking about people at these companies being interested in these models, I'd love to hear more about, like, in what way are they interested in them? How do they square this question of of user voice and deliberation with fiduciary obligations and and corporate ownership?
Speaker 6
4:45 – 4:45
If I can also jump in on that, abusing my unmutedness. I
Speaker 2
5:00 – 5:00
Please do.
Speaker 6
5:15 – 5:15
I also I'm curious if adding to that, if you see this as something that should be experimented with those shareholders. Like, are might there be a might there be ways of applying this process that include those people too?
Speaker 1
5:30 – 5:30
I think those are all excellent questions. I have a lot to say at all, and that would be its own three hour long talk. But I will I will attempt to to hone in. And I think, like, maybe the the first thing I'll start with is this challenge, which is how different do you actually like, when you really dig down into the details, how different really is the sort of the structure of a, let's say, a Facebook, and the structure of a government, and the set of incentives that it has, and the set of stakehold like like, shareholder type entities that it has. And and I think like when you the there are differences, but they're not as big as they might appear on first glance, and like just as a sort of like rough high level analogy you can maybe think of like the owners of real estate as, like, the shareholders of your nation. And, again, this is not quite right, but it gives you an intuition. And then, like, there's an executive, right, which is not, you know, there may be some democratic head to that in some extent. But other than that, they're they're they've got a whole bunch of legal issues of anything that they're gonna try to do. They have like, there's a set of, a set of similarities which are actually a a bit broader than they might appear on first glance. And then to to so just to get at the legal issues, I do see this as being part of, again, like, a set of of things that need to happen. And so one, like, key component here is we we should shift to a form of of of a more an economic system, which which isn't maybe always shareholder first, and, it's actually for an individual company. It's very easy to do. You can within The US, you can just have 50% vote and become a benefit corporation, and then you don't have these legal obligations in the same way. And then also if you there's an alternative frame for even looking at the existing economic system. This is the universal owner theory, and the basic idea here is just if you are a diversified shareholder, and most people are diversified shareholders, so you you want an index fund like the s and p five hundred, that's and your your pension fund does, and that's a majority of your wealth. It is in your best interest as a shareholder to not have one company do well in a way that hurts everyone else. And so, actually, it is in the shareholders' best interests not to have a a Facebook, like, screw over everything, else not to mention cause like literally civil war because that's really bad for your shares as a shareholder. So I think there there are both legal arguments There and there's sort of, you know, in this movement that's forming, and shareholder commons is one amazing organization that's working on this. I'm a huge fan. There's engine number one, which is doing this for for for climate. Again, just like this frame of, like, oh, okay. Even within our existing system, these these these differences are are actually smaller, and and the legal limits are maybe not quite as as as sharp as as they might appear. There is some case law, but it's again, you can come back as a corp to overcome it, and then there's a bunch of legal arguments that are being put forth that actually make that even you don't even need to become a benefit corp to to do that. But then going back to, like, the question of, like what about the just the internal incentives of these organizations? The I'm trying to remember the the specific I guess you're talking with just the the business model being in conflict with the outcomes of the assemblies. And yes that can be true that's also true for governments right like you know you can have an assembly that says you should spend money on a trillion things and then you're gonna bankrupt the country Right? Like, this is not unusual. This is this is, like, the these these things are are are similar across domains. There's a slight difference in terms of the intent and the purpose. The government exists in theory first to support the citizens and then to support its own wealth increase. So that that is a like, I don't wanna be, like, papering over that difference. But the the the idea of having a a mandate for a given representative deliberative process or, you know, a particular assembly, and that is a sort of that is the space which within which those recommendations are are or or even decisions that will be applied. That, that limits the extent to which you're gonna have maybe direct conflicts there. They can say whatever they want outside of that. Just like an assembly that's about one issue run for government can say every anything they want about another issue, but that isn't the mandate of an assembly. And so, like, and then those things, you know, do happen, and they'll happen in these other domains too, and that's okay. Sort of similar to how, you know, you have this judicial this sort of judicial body of a Facebook oversight board has a set of recommendations, and that's great. But that isn't necessarily the mandate of the organization to have those be directly implemented. And so you're being able to scope mandates to particular governance processes. You can you can do that however you'd like and I guess like one thing to to add here is the way that the the most compelling approach I've seen for implementing this at a societal scale like not just for not for platforms but for for governments is, by Terrell, and I'm gonna slaughter, his last name, a, but it's a the peeling model for it's appealing as when you peel off one domain of whatever the government is doing. Like, let's say, it could be platform regulation and say, well, okay. You're gonna have a platform regulator or or you're gonna have, like you're and that regulator is the the decisions of that regulator the the goals of that regulator, all that is gonna be defined by representative deliberative processes. And so and now that that's sort of out of the scope of the legislature, or it's at least not something the legislature is going to be tackling, you know, if it wants to give legitimacy to that new that new that new organization that's gonna be focused on that. And so I I can I see the same thing applying to platforms? There's a whole set of issues that they do not want to be making decisions on, and they wanna peel that off and devolve that to these third parties.
Speaker 2
5:45 – 5:45
Yeah. I I think what you just said at the end is really important. I I do have more to say, but I think the differences are are real between governments and and companies. And and I think there are some really interesting historical cases where people, you know, decades ago tried to implement platform democracy and then and then ran into legal issues that caused a reversion to autocracy. I would love to explore this in the three hour session. But I wanna turn to Bea, who who who says they're back on.
Speaker 6
6:00 – 6:00
I am. Hey there.
Speaker 2
6:15 – 6:15
Yes. We hear you. Great.
Speaker 6
6:30 – 6:30
Awesome. I I mean, one of the things I love about the MediGo community is that there are always ideas that challenge me to expand my idea of what's possible. And, Aviv, I know we've talked about this, but I feel like this is one of those things that, like, a part of me is like, that can never happen. And then, like, it's happened and it's happening, and that's really cool. So I wanted to know a little bit more about making it happen. You mentioned, you know, you're potentially piloting some things or in conversations with some, organizations that are exploring this. I wanna understand a little bit more about, like, what what's a good fit? What's a platform anyway? Like, for instance, you know, I use tools like Asana and Slack. I don't know if they count as platforms, but I could imagine that, you know, there's there may be some relationship between, like, user focus groups and platform assemblies. I don't know. Like, maybe tell me if that's, like, wrong and insulting. But, like Not
Speaker 1
6:45 – 6:45
at all. No. I think I I mean, it it yes. Yes and no, actually. It's basically yes. It is deeply insulting. And also, yes. There's a negative.
Speaker 6
7:00 – 7:00
Yeah. I just wanna know, like, I would love to see more of this be tried. And so, like, who are you trying to connect with? Like, what you know, when you say looking for more people at platforms who are interested in this, like, what does that mean to you?
Speaker 1
7:15 – 7:15
Yeah. Okay. So first, let me address the, like, connection to focus groups and Asana and whatever. Like, so you can one way to think about this, is, like, design thinking is in a way to understand what users want. And so like it's a set of processes that you can do to do that. This is also a way to do that, the difference here is that the agency is going toward the people being designed for and you're taking into account not just the goals of the system creators, but the the externalities, the broader sets of impacts on on that population. And so the I think that's the the big distinction here. So it is part of the spectrum of caring about the people who are being impacted and trying to support them and and caring about the the system being useful for people but it's it's a it's a it's just it's a broader approach and it's a way that gives far more agency to the users and yes you can have this for Slack for Asana I mean, I I would love to see agency being given to the the people being impacted by systems broadly. I just see the that is being, like, urgently, critically important for the these global platforms, that we really should not have a level of I mean, we we really need additional, like, processes to to to ensure that they can act, and do act in a legitimate and democratic way on the issues that are critical for our societal functioning and democracy to even exist. And so that that's the the piece. I think there was I forgot the other part of the question. Oh, which is what what was what when who are the people? What do they need to do? So, like, at least within platforms, there's sort of a a few different types of entity that might care about this. So you have product leaders who are like, I'm man I'm owning this product. They might even just be a product manager or maybe a product manager manager or, you know, or VP of product, you know, all the way up and down the chain. Just like, I have a product. I don't know what to do about this thing that has, like, these broader impacts beyond the product itself. What do I do? How do I think about this? Anyone who asks that question, this is a thing that helps support that work. And it's something that is, you know, not all that expensive to do, at least and I mean, at different scales, definitely different levels of expense. But compared to the impact the cost of doing the thing wrong or a thing that in a way that really pisses off stakeholders doesn't have that legitimacy, it can be a significant net win in terms of, like, the total cost for that platform. Obviously, because of the way that accounting works and sort of hiring and this and that works, the actual organizational types of platforms, the product manager may not realize the benefits of that. And so this is where policy people come in because they're the ones who pay the costs. And so the the other type of stakeholder that's often, like, a good one to to sort of kick this off of in a platform is someone who's doing policy, you know, especially maybe within a European context, because they're actually doing things. And and I mean, a number of other of other, regions also, but that's sort of the most obvious where they're like, okay. We're getting pushed in all sides here. We wanna have we wanna elevate this debate to a higher level. Can we partner with the regulators even, to to actually run an assembly on the impacts of our platform in this particular area, and how we should be doing this better. And so I think there's like a huge benefit to those people and so they're also excited about this. I think there's one other constituency that I can't quite remember right now, but those are the two that immediately come to mind.
Speaker 2
7:30 – 7:30
Well, we need to wrap up, so, the forgetting is convenient. But thank you so much for a great discussion. We can continue it in the Slack. Let's take a moment before we leave to, unmute and applaud if you're able. Share appreciation to Vivian. Three, two, one, go.
Speaker 1
7:45 – 7:45
Thanks, Sal.
Speaker 2
8:00 – 8:00
Thank you so much for for sharing all this work with us, and we, you know, look forward to to continuing the the conversation.
Speaker 1
8:15 – 8:15
Yeah. And if folks wanna stick around, happy to there's a few things I'd love to sort of hear perspectives on myself.