Keynote: ⿻ Infinite Diversity in Infinite Combinations by Audrey Tang | Devcon SEA

[Music] really excited uh to be here uh I missed the bota one I was in the Osaka one so happy to uh be back at Defcon um yeah thank you and today's talk uh is plurality uh infinite diversity in infinite combinations uh you can read all about it at plural. net which is a public domain book now uh as you have seen in the trailer the good end of ancestor back in 2014 just a decade ago Taiwan was facing a real public trust problem uh the trust in the Democratic Administration was as high as 9% which was not very high uh and the society was very polarized and very angry energies everywhere and so the point uh that we made back then was that we should not succumb to the anger and outrage that swept across much of the world at a time and instead we should turn conflict into co-creation and there's a saying that says to give no trust is to get no trust and in the span of just six years we had rebuil the trust by radically trusting the people and so by 2020 the public trust has risen to over 70% and so this is the journey that began a decade ago the sunflower movement March 2014 uh was when people were really deeply unhappy about how the administration and the Parliament was handling a particular trade negotiation but what said is movement apart is that we said that we're not protesters protesters is against something we're demonstrators demonstrators build Alternatives so we occupied peacefully the parliament for three weeks with half a million people on the street and many more online and the occupied streets around the parliament became a live open citizen assembly anybody can come there and learn about how exactly does the trade agreement affect them and they can engage in a facilitated conversation both in person and online and every day we reach a little bit consensus about those specific issues and read them allowed and after three weeks something very rare happened in the history of occupies the speaker of the parliament accepted or our demands and so that was extraordinary and by the end of 2014 the central government invited us the Civic activists to join the cabinet as reverse mentors advising the cabinets in art of how to rebuild public trust and so the idea the key idea here is bridging systems the first thing we did in 2015 was building such bridging systems online platforms that mirror the facilitated dialogues that we had during the occupation for example this is a real screenshot back in 2015 about Uber which involved nonprofessional drivers uh and which uh really offended the taxi drivers uh caused a lot of conflict both in the street and online but instead of letting the discussion devolve into outrage on social media we build our own pro-social media where citizens can express their views constructively and this is the polce platform it's an open source platform where people can propose ideas like insurance is very important for right sharing and so on and as you can see there's this agree button and disagree button but there's no reply button with with no reply button there's no room for trolls to grow and uh you can see as you click agree or disagree that your avatar that blue circle moves toward to a cluster a c's cluster uh where people share with your views and you can see what bonds you together and what are dividing the groups and the idea here is that we focus on the bridging idea the more an idea Bridge across groups the more bonus it gets and so only the idea that speak to all the different groups can get agenda setting power to be part of the policy agenda and so during the Uber deliberation people first agree on the basic facts like what's actually happening and then we ask how do you feel about those facts and some were angry some were you know happy but everybody is an expert in their own feelings and this allowed us to Cluster based on the shared sentiments and so by the third week we had a set of very bridging ideas like search pricing is acceptable while the demand is high but undercutting meters is not and so the final regulations setting this as agenda ensure fairness for both Uber and the local cooperatives in rural areas and so on and so the lesson we learned back in 2015 was that polarization is not a function of the people involved it is a function of the space and so we can Design pro-social Systems that reduce polarization as a function of the space making the group informed consensus into a common knowledge like a group selfie that everybody sees and say oh actually we do agree with most of our neighbors most of the time on most of the points and it's just one or two things that divides us and that group selfie is very powerful now the idea has since then influenced the mainstream platforms uh such as Twitter now x.com uh and also YouTube so maybe some of you have tried the jury duty that is the community notes uh it follows the same principle so if only one side on the left on the right agree with a note it does not get featured it only gets featured if it Bridges perspective from both sides encouraging more and more balanced contributions and so by promoting those notes that bridge differences the system incentivized this idea of constructive engagement and so more generally while the social media connects us globally it often treats the individuals as isolated units and prioritizing that engagement that do hit over genuine connection and so instead of these efforts to exploit our social fabric rather than regenerating it uh we can do better and the poce community notes are just prototypes to what we call pro-social systems and the main idea of prosocial systems or plural systems as vitalik wrote in his blog is to treat the connections between individuals as first class objects and that means designing platforms that recognize and enhance the Relationships by naming the relationships and featuring them as part of the group selfie and so transparency is really key because many platforms already understand the networks that we belong to and the interest that we share but the information is concealed from Individual users and exploited from personalized targeted advertising and so by openly labeling those content with Community it connect us to whether cultural professional or social we can empower the people to engage much more meaningfully and so this idea has many applications and in the book we discuss a lot of them but basically all of them falls into this uh tradeoff this production um possibility Frontier as you can see on one side deep collaboration happens in small in homogeneous groups while large scale interactions tend to be more superficial like just capitalism uh but what if we can push the frontier forward and so that is the main research agenda for plural systems what if we can Design Systems that allows for the same deep meaningful collaboration but across larger groups or we keep the same skill of people engaged but much more meaningful much more higher bandwidth democracy and so by designing spaces that accommodates this kind of infinite combinations uh in infinite diversity we can then move toward a much more interconnected society and tackling more and more emerging challenges so I'll just use one example which is the middle uh the augmented deliberation uh more recently uh we tackled in Taiwan uh this March a society-wide issue which is a emerging issue called defects and other synthetic content on social media and so this Mark we send out 200,000 random text messages SMS to Citizens across Taiwan from 111 which is a well-known number that says this comes from the government so you know it's not scam and it asked a very simple question how do you feel about information Integrity online and from those who responded uh who said okay I'm willing to engage in online augmented deliberation we randomly selected 450 people through random stratifi assembly or sortition a microcosm of our society in terms of age occupation region so on so forth and so this microcosm were divided into 45 online rooms with 10 people each and instead of professional facilitators the room now autof facilitates the AI power tools guides the discussions without a human facilitator and the platform transcribes the real time uh transcripts conversations key points points of disagreement points of agreement and even encourage the quieter participant to share their thoughts more and so the AI assist deliberation enable us to scale meaningful conversations in real time and synthesize different ideas from those 45 different rooms into a coherent bridging idea that we then read out to the 450 people during the plenary for example one consensus that was Wen this way was that social media platforms need to be liable if they fail to verify the mandatory digital signatures of online advertisements in other words if um an advertisement shows awn selling you some scam or something and the platform did not secure a digital signature from me and somebody get con for $1 million then that platform should be liable for $1 million and so we took those Collective insights and within just three months just passed the legislation with the support of all the different parties and so the new law requires social media to ensure the authenticity and all advertisement so it's not content level moderation or censorship but rather actor level Providence and so this process demonstrated how technology can enhance Democratic participation turning this complex Challenge and basically deliberate in the same speed as the technology evolves and so uh our journey in Taiwan taught that uh if we instead of broadcasting just broad listening uh from the society we can just integrate those very different conflicts polarized sides and turn them into co-creation it's not just limited to the time one uh for example in Japan this year a young political newcomer a AI researcher name is takahiro anos uh embraced his principles after reading the plurality book uh and run for the Tokyo governor and so inspired by the plurality book anoong um basically used AI to um deep fake himself uh so uh he live streamed as a vtuber for 17 days answering almost 10,000 questions from voters you can call into a hotline have a conversation with his voice clone uh but the point is that it's all open source and his Manifesto his platform is on GitHub and so this kind of collaborative co-creation again show the people in Tokyo like in real time what their preferred platforms are and so ano's experience um basically illustrates how the principle of broad listening can result in a very high quality platform independently rated as the best platform even though he did not win uh Co Cas on one uh but the idea is that we can enrich our Collective experiences and with a stronger more interconnected social Fabric and so that leads to the idea of pro social media what you're seeing here is a prototype uh called cortico thatai from MIT that exe exemplify the idea of pro- social media uh and the software is to enable small Gatherings like religious community educational settings to hold meaningful listening sessions and the software again transcribes and anonymize the conversations highlighting the points where people resonate and share those resonating um ideas uh across the bridging groups and so this creates a network of understanding where insights from one group inform discussions in another group and ultimately a collective sense making report is done again with the help of language models and then reflecting the collective voice of the people so that people can then understand deeply uh what they all care about and the idea of pro-social media then regenerate the social fabric rather than maximize individual engagement and this can also offer a new Revenue Source uh because for communities and also for personal subscriptions we all want to connect more meaningfully with our communities and so uh this leads to the real world um application part uh so in our Pursuit uh to do more pro-social digital environment uh this year there's also a very interesting critical opportunity so one of the world's most influential social networks is called Tik Tok uh and we think rather than resorting to bonds we're putting together what's called a people's bid to acquire the US assets of Tik Tok and turned it into exactly that sort of pro-social constructive Community Driven space uh as you can see uh Tik tok's recommendation algorithm has been a double-edged sort last year there's research uh that shows um a average undergrad in the US if you want them to stop using Tik Tok uh you have toay pay them $60 or so a month uh for the utility that's lost but if there is a way for uh them to make everybody around them including them themselves quit Tik Tok they're willing to pay like 30 uh US dollars per month uh and so this is a kind of collective trap situation where people find it very difficult to leave due to the network effects and peer pressure and so the people's bid basically seeks to acquire Tik toac us and shift it into the open decentralized infrastructure not unlike forecaster or lens and allowing users to Port their social graphs and choose their own experiences so there's no um immediate loose if you migrate away from one particular front end you can keep all the values in the network and so by replacing the existing recommendation engines with bridg making algorithms we hope to promote content that Fosters understanding collaboration across those diverse different viewpoints and research from Stanford already show showed that such bridg makak algorithms can maintain the same user engagement while enhancing the quality of interactions and overall mental wellbe so if you know someone from the web3 space who would like to architect Tik tok's eventual migration to the decentralized social networking protocol like a CTO for the people's bit uh please do get in touch so uh now before moving to the Q&A uh with 10 minutes left uh I'd like to share something personal um back in 2016 when I was first named digital minister in Taiwan Taiwan did not have a digital Minister before and so the EDR asked me would you like to write a job description to uh describe what the digital Minister do and uh in Taiwan it just so happens sh way means both digital and plural so in addition to being a digital Minister I'm also a minister of plurality uh and so what follows is really a prayer it is my job description as digital Minister and now cyber ambassador of Taiwan goes like this when we see the internet of things let's make it an internet of beings when we see virtual reality let's make it a shared reality when we see machine learning let's make it Collaborative Learning when we see user experience let's make it about human experience and whenever we hear that a singularity is near that a just always remember the plurality is here thank you so [Applause] much any questions thoughts ideas comments well feel free to um get in touch I'm at Audrey t on most platforms thank you oh there's a question there sorry yes please uh thank you really great talk um yes I was just really interested with the experiments that you've run like if you ever have trouble with people engaging and if you've explored different kind of incentive mechanisms to get people involved in all this great work and research yeah that's a great question yeah so the key to get people engaged uh is fast fair and fun uh fast meaning that you need to offer a kind of instant gratification so for example in Taiwan even for people under 18 anyone who secur the signature online for 5,000 people can kind of force a ministerial response within just 60 days at most and so it encouraged people uh like according to the iccs the international citizenship uh and civic education standard the Taiwanese 15 years olds are top of the world when it comes to feeling that they have the agency that they can change the society and we publicize uh the results which could be fun you know one of the results early on uh was Manning plastic straws from the bubble T takeouts the Boba takeouts that was very successful started by the 16-year-old by the time she was 17 she became a cabinet level uh reverse Mentor there's also another very popular petition uh that says um we would like to go to school one hour later because study shows that one more hour of sleep actually as opposed to one more hour of study actually increased the grades uh and the Ministry of Education was like yeah that's a great idea and so that was passed uh and included into law and so by publicizing yeah it really is science one more hour sleep does help study um so by making sure that we publicize those early wins those quick wins we make sure that people who have like otherwise longer term you know planning thresholds and so on can get the support like a 16 year old and 17 year old they can get a support from 60 years old and 70 years old because both have more time on their hands and care more about the long-term sustainability anyway and so each even partial or failed petition then builds the community that pushes the next petition forward I hope to answer a question you great any other questions um there's quite a few there thank you thank you for the talk I'm huge fan of your work H thank you for that oh yeah sorry for the for the mutex sem we can ask coordination questions yes yeah please okay uh so again I'm a huge fan of your work thank you for that um but um while we try to use what you suggest in in web three my question is what would be your call to action like a waiter wer call to action for all of us also outside web three because it seems like it needs like a huge leap from where we are to those new systems that can you know stop the World from fing apart thank you great um the most um interesting experiments early on I saw in quadratic voting quadratic funding this kind of high band with plural systems they all started in web3 space and mostly from the ethereum ecosystem basically what we use in Taiwan either for the presidential hackathon or quadratic funding would basically look at uh the ethereum ecosystem see what worked uh and then took what worked uh and in the real world applications so although I guess it's not strictly speaking real world application of eum is real world application with ethereum uh and uh nowadays I think with the new found superpowers of ethereum to settle the transactions and so on so quickly um I am thinking that most people who want to build such pro-social system who want the social graph to persist uh in a way way that transcends individual experiences and so on they now see like what we just saw from the Sony uh and the Sony L2 idea we can take existing um traditional institutions centralized institutions like Tik Tok uh and then offer some sort of a ramp so that it become uh decentralized and then bridging into the existing very exciting decentralized social systems so I think that sort of um conversion the web 3 to um like Web Two ramp uh and offering a kind of a easy ramp for the web to people to come aboard I think that is one very exciting opportunity that I can see but keep doing the experiments of the bridging bonuses and so on in the web three world because after all for a policy maker like pointing to anywhere in the world where something worked is a sure way to get it into the standard and Consciousness uh of people so keep up the good work uh question do you uh uh see the innovations that uh you've been working on in Taiwan on on Democratic systems are they merely a better way of manufacturing consent or is it possible that at some point we evolve beyond the need for hierarchical structures um so a direct democracy could actually lead to better uh decisions within the right coordinating structures that's a great question um indeed uh I think currently the examples I gave is more like in design thinking terms it's like in the first diamond it is about um discovering the common values and defining how those values should be applied to solve emerging issue uh but in existing representative democracy the second Diamond which is the development of the budgets and policies and the delivery of such things in the execution layer in the administration uh these stay the same so basically this is like a add on uh to the existing representative democracy but we do also see for example during the pandemic or the infodemic the disinformation crisis during the pandemic we do see that there are certain functions that people in Taiwan do not want the government to run for example factchecking um they would think that if you centralize the authority of factchecking into something that's run by the government then the government has an incentive of basically uh saying uh things that are front of opposition party as as falce or something so in Taiwan we do not actually concentrate that power into the administration but rather rely on the uh collaborative factchecking in the Civil Society on the open source ecosystem so that for example Facebook or Instagram uh The Meta ecosystem they partner with the time fact checking Center or the mopen or the collaborative fact checking system and these are in the social sector these are in the Civil Society and we make sure in the government that we just offer like very quick responses to the fact Checkers instead of assuming that function ourselves so I think the more um practical way to think about it is to feel uh in a society whether there are some functions that people just do not want the government to centralize and from those fields we can then use the same coordination measures and then extend to the development and the delivery um layers and once people have good Peak experience uh like the uh Society wide deliberation uh and then we can piggy back on those experience and then expand more for direct democracy or participatory democracy but it needs to be done peace meal instead of just assuming all the governmental functions before the Grassroots the Civic muscles are ready thank you great question yes hey um so so my question kind of connects to direct democracy MH uh like the last question and uh if if I had to make an analogy on the um on your kind of um SP and idea is that we're trying to migrate realities so we trying to migrate from a physical reality to something that is more virtual governed by bits and bites maybe later other things mhm um but physical reality has a fabric which is very basic does one person one vote and that kind of governs morality in some way right um so how are you trying to migrate that because physical reality has Bad actors and Bad actors can exploit the lack of fabric that you have that you don't have in virtual um realities namely it's called civil um stability or anything like that so what I'm curious to understand what you're thinking about that so just uh to clarify the question the question is that in physical reality like each person is physically in one space uh it is impossible for me to Fork into like 10,000 clones or robotic clones uh and it has been like that for thousands of years so I'm that's that's my question basically yeah how do you migrate that principle which is physical to something that is virtual yeah um so for example uh in the Pol system or the systems that we run we of course need to ensure that each person only have one presence uh in that system uh we usually use SMS which works domestically because it's if you try to get like 5,000 SMS numbers the in money launching people discover you very quickly uh but beyond the borders uh it is uh then the need for some sort of interoperable like personhood credentials the ways to kind of disclose that you're not a bot without overly disclosing anything about yourself that becomes very important um and so uh just um I think a couple days ago Finland uh is now running a hackathon uh that tries to solve this sort of problem because they want to use polce as their society-wide public digital infrastructure and so they don't want like fins overseas uh to have a lot of trouble authenticating but they they also don't want uh people being able to impersonate 10,000 people and so they used exactly the same toolkits that you're all developing here the uh zero knowledge proofs The Selective disclosure verifiable credentials and things like that and so it's not to make the virtual world a surveilled world at or world where it's like real name transactions but rather just to preserve one single feature from the physical world which is one person can only be physically in one room uh at a time uh so that is the feature that we're trying to um imitate uh the in the virtual world without uh making it into a surveillance of capitalism space so I I'm not exactly sure that to completely address your question but that is like the number one uh issue we face when we deploy that in the real world any other questions uh yes one more question okay hello um and thank you for for your talk it's inspirational quite um if I can put this question together I'd ask uh what's the biggest threat from State actors um against radical change and Civic movement like what do you think we can do about um governments trying to subvert their own Societies in the way that you've spoken yeah I think the the number one threat that I see uh is this sense of the society being so polarized that any sort of organization or mobilization or trust uh is impossible so it is in a authoritarian government's best interest to make sure that each individual feel like a individual uh and not at all horizontally connected not at all participating in anything that's meaningful and change making uh and so to counter that a lot of the work that we've been doing is to ensure that even at a very small community Grassroots level there are ways to quickly see that we do actually agree on common actions and can make those common actions together indeed back in 2014 uh in the occupi parliament in Taiwan all the different Corners in Taiwan were occupied by different NOS the Civil Society organizations that has been around even when Taiwan was a under the martial law was a dictatorship and they did not have the um canvas to tackle like overtly political topics back then maybe they tackled like consumer right protection or Environmental Protection or Collective buying like cooperatives and so on but after a few decades the relationship most Muses and Civic muscles grew to a point where people do feel that they're in it together and then uh they affected a peaceful transition to democracy in Taiwan and also ensure that it's not violent because at any given time people belong to intersection groups of Civil Society actors that they can feel that even though maybe the president wasn't yet directly elected that didn't happen until six uh 96 but throughout the 70s and 80s more and more room gets opened by the Civil Society actors so to counter the pessimistic polarization or manufactur polarization the easiest way is just to get organized and just trust each other even on a small scale thank you all right thank you so much for the great questions live long and prospering thank you thank you Audrey thank you AUD amazing talk okay with this we actually are concluding the programming for this main stage uh at this point the stage will be closing down it will actually be dividing into three stages which will reopen in the afternoon for more programming uh starting in 24 minutes are more talks in other stages uh around here on the same level so you have talks around uh yeah introduction to multilateral Trade Credit set off an PC onchain Capital allocation Workshop what's going on in into the pectra upgrade by Christine and bridging trafi and defi through ethereum and evm a pathway to Innovation so check out those talks in those other stages we'll need everyone to clear out of this stage so we can reset uh and thank you all for joining us here this morning [Music] [Music]