Your Undivided Attention - Digital Democracy is Within Reach with Audrey Tang (Rerun)

Episode Date: March 24, 2022

[This episode originally aired on July 23rd, 2020.] Imagine a world where every country has a digital minister and technologically-enabled legislative bodies. Votes are completely transparent and audio and video of all conversations between lawmakers and lobbyists are available to the public immediately. Conspiracy theories are acted upon within two hours and replaced by humorous videos that clarify the truth. Imagine that expressing outrage about your local political environment turned into a participatory process where you were invited to solve that problem and even entered into a face to face group workshop. Does that sound impossible? It’s ambitious and optimistic, but that's everything that our guest this episode, Audrey Tang, digital minister of Taiwan, has been working on in her own country for many years. Audrey’s path into public service began in 2014 with her participation in the Sunflower Movement, a student-led protest in Taiwan’s parliamentary building, and she’s been building on that experience ever since, leading her country into a future of truly participatory digital democracy. 

Hey, everyone, it's Tristan. Lately on Your Undivided Attention, we've been talking about democracy, or rather the failures of democracy in the digital world. And so this week, we wanted to offer a hopeful and working vision of a digital democratic future. We wanted to rerun the story of Taiwan, as told in our conversation with Taiwan's digital minister, Audrey Tang. This episode is from 2020, but it's as relevant today as ever. Because oftentimes in the social media conversation, it seems like our

interventions and solutions and regulation are just trying to get to a world with maybe 10 or 15% less toxic social media. But is a slightly less toxic social media living on top of a democracy going to enable hopeful visions for how democracies can work in the 21st century? What we need is a vision where technology plus democracy equals stronger democracy, not just slightly less bad dysfunction. So with that, here we go. Imagine it's January 2021, and the United States and Western states around the world decided to become 21st century digital democracies. The United States hired its first digital minister.

We went through sweeping reforms that entailed a modernization of Congress, transparency of each member's votes, video of all the conversations that Congress members had with other lobbyists and guests. Imagine a world where conspiracy theories were all acted upon within two hours and replaced by humorous videos that actually clarified what was true. Imagine that expressing outrage about your local political environment turned into a participatory process where you were invited to solve that problem

and even entered into a face-to-face group workshop, that pothole in the street that's been there for four years, suddenly it's changed. Does that sound impossible? Well, it's ambitious and optimistic, but that's everything that our guest, Audrey Tang, digital minister of Taiwan, has been working on for her own country

over the last six years. I'm putting into practice the ideas that I learned when I was 15 years old, and that's rough consensus, civic participation, and radical transparency. Audrey Tang's path into public service began shortly after student protesters in Taiwan stormed into the nation's parliamentary building in 2014 and refused to leave until the government heard their calls for greater transparency. There was just one problem. The protesters had no Wi-Fi.

Enter Audrey, who walked in with a backpack full of Ethernet cables and installed a system of communication that was radically open, transparent, and responsive to the public's demands. I personally brought like 350 meters of Ethernet cables to make sure that the truth spreads faster than rumors that we live streamed. What's happening in the occupied parliament to a large project on the street where they can see in real time what's being debated and what people said in the occupied parliament. It was so successful that Taiwan's government invited her to build a system of communication for the country that Audrey calls a listening society. And surprisingly, it's working pretty well and it's transformed our society.

It's also important to recognize that this digital governance process had Taiwan be one of the most effective countries in dealing with the coronavirus. They never even had to shut down any of their restaurants. They never went on lockdown. When it came to misinformation, they actually were able to combat rumors and conspiracy theories about the unavailability of masks with digital tools that provided transparency about exactly how many masks were available. So when we look at an example like Taiwan, it's actually a model for democracies around the world. Maybe that's what the world could look like in the future. We can look at an example like this and say, oh, we could never do that here in the United

States. And what if now was the time to change that? What if this was the moment when we actually show that there actually is an alternative to the rising tide of Chinese digital authoritarianism? What if there is a version of digital democracy that actually is listening to each citizen? and that our news and technology feeds sort for unlikely consensus and agreement instead for outrage and division. What if we had an information environment that actually reflected truth?

That world is possible, and Audrey's work shows that it is. I'm Tristan Harris, and this is your undivided attention. Audrey, thank you for coming on the podcast. Hi. Hello, everyone. Hello, world. Really happy to be here. We have a slightly different format for this interview. I've invited Glenn Weil, who advises Microsoft's leadership on geopolitics and macroeconomics, and is also the founder of the Radical Markets, sort of Movement, Radical Exchange Foundation, to interview Audrey Tang. Audrey, why don't you take us into your personal story?

I'm trying to just even imagine what it was like to be there at the Congress. Did you sort of slam through the doors and kick it in and open your laptop and invite people into discuss online? We did both. So I was there the night before they stormed into the parliament. At the time, the members of the parliament were refusing to deliberate substantially a cross-straight service and trade agreement or the CSSA. And because of that, the students, along with around 20 different MPOs, occupied the parliament. The legitimacy theory is simply that the MPs were on strike, so to speak. And because of that, everybody needs to just take their seat and do their work because they refuse to do their work.

And so for about three weeks, totally nonviolently, people just talk about all the different aspects of the CSSTA, including actually the critical issue that other nations are gripling with now of whether we will allow people, That's People's Republic of China government components in the 4G infrastructure back then in 2014. And so 20 MPOs talk about 20 different aspects and we make sure that we live stream all the meetings. I personally brought a lot of Ethernet cables to make sure that we live stream what's happening in the occupied parliament to a large projector on the street. And even with the stenographers working in the Occupy Parliament, people who don't hear very well

on the very crowded street can still look into their phone on the kind of an IRC chat room where they can see in real time what's being debated and what people said. We make sure that we get the rough sense of the crowd every day and every day we push the consensus a little bit more so that by the end of the three weeks people have a very strong consensus about what to do, including not allowing PRC components to 4G infrastructure. So I would say it's a technologically amplified occupied, instead of like many occupied places. There's essentially no coordination between the media people who talk to the online world and the activists who basically hold the fort in the Sunflower Occupy, these two groups

work very well together, very closely together. And many people who are online supported the very few people who occupy the parliament. So when the police eventually came, they counter surrounded the police, making sure that the police cannot just evacuate the students. So, Audrey, one thing I see as essential to the approach you've taken is. is that the large-scale mass protests, the technology sector, there seems to have been a real integration of those very contrasting sensibilities in the way that you and the GovZero movement approach things.

A lot of people here in the United States, when I tell them your story, say that, you know, well, Taiwan is just a totally different society, you know. But I'm really curious what you think changed in Taiwanese society, how that change was possible, and what lessons you see for that? Yeah, if you ask a random person, I think, in 2010, on the street of Taiwan, whether people who are social activists, people who are civic media, and people who work on free software, will somehow unite under the banner of God Zero. People will look at you like you're crazy. So this is definitely something that really takes a special moment in history. But just like the relationship between research and development, once the idea is there, it spreads very quickly.

So the core idea of Gov0 is G0V. The idea is that back in 2012, a few friends of mine registered this domain name, G0V.TW. The core idea is essentially taking all the websites that they don't like in the government and make a fork that's to say a alternative based on civic technology. So all it takes is a letter change in your browser bar, G0V, and you get into the shadow government that works better. more participatory. So it's a very interesting idea. This is actually a critical point to contextualize for people.

This would be almost like there's Whitehouse.gov. Because Whitehouse.gov, let's say, host government documents or bills or executive orders. And here you are coming along with your civic hacker friends and saying, you know what, we think we can build a better civic tech infrastructure than this existing domain. And so you assembling those tools would be like building this sort of White House stuff. G0V. G0V. Yeah, exactly. And the idea that prompted this idea was a controversial government advertisement in 2012

that says, instead of wasting time deliberating policies, let's just get things done. Basically, authoritarian advertisement. I think it's one of the first YouTube advertisement that the government ever filmed, and it was also flagged as spam, meaning that people don't really like it. And so these civic hackers build a citizen auditing system for Taiwan's central government budget, essentially saying that the reason why the Taiwanese people cannot make sense of the budget is not because we're dumb, it's because there's lack of visualization. There's like 500 pages of PDF files, and of course nobody can make sense of that.

And because they made the data from the accounting statistics of it accessible, easy to understand and interactive the public could then rate and comment on every item in the budget. What tools did you have to string together at that time? Because I think that helps set the stage for people's understanding. Basically, the productivity software that we all are used to, including collaborative editing documents, collaboratively editing spreadsheets, real-time chat, but they're all open source, and so that's some of the earlier infrastructure. Of course, it's now part of our basic participation platform, which has more than 10 million

people out of 23 million people. But at that time, it was just a very few people doing this kind of forking the government work. I mean, they could have tried to say start some Facebook groups online. and created a governance system through Facebook. And yet, obviously, you didn't do that. It sounds like you built your own tools for spreadsheets, your own up-down voting agreement, consensus.

I like this, I don't like this. I agree. I want that policy passed. You know, you could have used some of the existing corporate things. Why? And what was the nature of doing it publicly? Well, first of all, of course, we want to make sure that everybody can fork our own code.

That is the idea, open innovation, making sure that everybody who have a different idea can just fork their version of that. civic tech service and then build their own reimagination of it is very powerful because then you get to merge the best or at least a better practices without get into ideological conflicts. And so basically proprietary software ensures a governance mechanism that is very arbitrary, basically the first mover, the first person who built that ecosystem get to ultimately decide whether they're a benevolent or non-benevolent dictator. But if it's free software, and the benevolent dictator is always put into quotes

because everyone knows that you can always fork it. And that kind of forces the project leader to be more inclusive and also allows for more diversity. So you used all these technologies to form this coherent set of demands. And then what was the moment, what was the process by which the government, at the time it was the nationalist Kuomintang government,

which was generally quite conservative, certainly not necessarily the most open of all possible governments, what was it that brought them around to recognizing these demands and not only recognizing the demands, but also recognizing the efficacy of the way that they had been presented? Definitely, it's the election of 2014, which was a mayoral election. There's at least two mayors who explicitly run on an open government platform, and there's many other mayor candidates that didn't quite say open government is my number one priority, but they at least supported open government, radical transparency, and also the Occupy Movement.

So the landscape changed because everybody who did not support the occupied efforts lost the election, sometimes very surprising to themselves. And everybody who occupied the Parliament, or at least helped occupying the Parliament won the seats in the mayor without preparing the inauguration speech in at least one case. So that's basically a new political will expressed by the citizen. And so through that new election, you also became digital minister? Not at that time. So I was a reverse mentor, a young mentor, to the then minister for law, also for cyberspace law, Jacqueline Tsai.

So you started basically teaching them how to use these different tools for building consensus. Yeah, yeah, that's right. That's right. Yeah, I personally trained more than 1,000 career public servants. I'm just imagining the parallels here of what that would be like in the United States. It would be as if you came to the United States government and started hand training, you know, thousands of government representatives, et cetera, how to use different open source tools to modernize Congress,

to host debates online to be able to put up proposals in a system where people could comment and make suggestions about what policies that they wanted to reform our most urgent problems. It's very hard to envision that. And yet that's, I think, what is also a lesson from your work that, you know, here you are doing this work on the outside of government that was out competing some of the inter-coordination collaboration tools that existed. And then you beginning to reverse mentor until you're ultimately brought in to lead as digital minister of Taiwan.

Yeah, and there's always an outside game. I must say that it's always important that if somehow I waver from this direction or the cabinet waivers from this direction, that there's always this implicit threat that the Congress can get occupied again. It's a remarkable story, Audrey. I'd now like to go into some of the specific tools that you talked about. Some of the ones I've been particularly impressed with, wiki survey tool, pole. I.S. Could you just talk a little bit about how that works and the role that it's played in helping to reach consensus in Taiwan? So the user interface is very simple. You look at one sentiment from your fellow citizen and you basically click only upvote or download or pass. There's no

reply button, so there's no place for troll to grow. And you also see your friends and and families on the various different sides, so that they're still your friends and family. And so it shows both where you stand in relation of these ideas, because each yes or no sentiment is an extra dimension. And so just imagine a n dimensional space, where n is the number of sentiments shared by people.

And then the algorithm automatically do a principal component analysis and visualize that in two dimensions what's the most contentious points are. And also they did a K-E means clustering showing the people with shared sentiments and what unites them. And so at the beginning, people were on all the corners, but because we say we only give agenda-setting power of our face-to-face consultation to anything that people can propose that can convince a supermajority across all groups. And so they converge on feelings that resonate not only with their aisle, not only with the like-minded people, but across the aisle. So instead of distracting, the software automatically attracts consensus. And so after we get a set of feelings that resonant with practically everybody,

then we get the stakeholders on the same table, live stream that meeting, and basically ask each them, here is the consensus, the six consensus of the people. Do you agree? If you do agree, how do we translate that into the regulation? And they are bound to the words that they said during the live stream consultations. And so that's how we ratify the agreements a year or so. later in August 2016. And everybody knew it's coming, everybody anticipated it.

Do you have an example of a problem that ultimately found an unlikely consensus? Because what I love about the way that you talk about it and having designed Polis, I guess you'd call it, is it sorts for unlikely consensus. So unlike our current social media, which sorts for what's most outrageous, which is usually what's most divisive,

you're sorting for the opposite. And I think an example to help people understand how that solved that problem. Sure. So for example, we just run, a few polis conversations around how to make contact tracing easier. So, for example, one of the consensus was a health information recorder on your phone. It solves the dilemma that most patients may not recall their whereabouts easily,

and when the medical offices come to them, they may actually accidentally divulge more information about their friends and families, more private details than the contact tracers need to work with. So instead, an app would generate a one-time-used link and a dashboard that gives them minimally required information for the medical officer to do their work without compromising their friends and families' privacy. It doesn't use Bluetooth or indeed any transmitting technology. It's because of that people would be much more willing to use that sort of tools.

So one other thing I've been really fascinated with is the actual process that helps stimulate and create the initial proposals that go into this, which is these idea of data coalitions. And I'd love to hear a little bit more. about how those operate and how they contrast with other sort of data control regimes that exist in other places in the world, because that's obviously a challenge that many societies have struggled with. Definitely. So in Taiwan, back in 2014-ish, people were suffering quite a bit, especially in the mid and southern cities, PM2.5, that's air pollution. And there's a lot of different theories, whether it comes from transportation or from coal-burning plant, or maybe it

comes from overseas from across the Taiwan Strait. There's lack of precise data. And so the GovZero, the GovZero Air Pollution Observation Network, working with the Airbox community. Airbox is a very simple air quality sensor, which was becoming very popular. It's less than $100.00. So all interested people, many of them primary school teachers, teaching data competence, can partner. by providing real-time air quality information on their balcony, on their school or in the office. So little by little and bit by bit, thousands of contributors, nowadays hundreds of thousands of contribute, accumulated a diverse network of at least tens of thousands of sensors, which is closer to the people and provides real-time air quality in the actual places where people are active.

That's an enormous amount of environmental data. We put it on the National Center of High-Speed Computation. We make sure that air products, meteorology, water resource, earthquake, disaster relief, and so on, are integrated into this in-place computing environment. So all primary schoolers or high schoolers, as part of Slansphere maybe, can write some code that very quickly run on those outside data to discover correlations between the social activities and environmental phenomena. Wow, it would be like starting a forked website in the United States for the EPA.gov and then starting to have civic sensors for climate change, floods.

climate, methane, signals, et cetera, and then starting to let people actually contribute their own sensors, just to, I think, contextualize it for people. It's fascinating. And because of this infrastructure massively lowered costs of any future innovations. So the data collision of the Airbox community, for example, negotiated with the private sector and the public sector saying, you know, we would be happy to provide our data for evidence-based policymaking. We will work with you to calibrate our sensors against the humidity, but in exchange, we demand that in the industrial areas which we cannot get to, like industrial parks, the lamps, which is owned by the public sector, need to also install those microsensors that we can complete the puzzle together and also

prompted new presidential hackathon teams, such as the water box, which does the same, as I mentioned, for the aquacolans, where they put this water pollution sensor in the riverways. And then, I don't know if you want to go here yet, but you did something. very similar for coronavirus, as I understood, in terms of understanding masks supply. Because I think the theme throughout your work is people always throw around that transparency creates trust, but in both cases, you have this sort of notion that having transparent numbers as contributed by citizens of the air quality measures in their own city where citizens actually are similarly for masks.

Do you want to talk about that? Sure thing. So mask in Taiwan is a social signal that remind people who are wearing them to not touch their own face with unwashed hands. and so they need to wash their hands much more vigorously. And that's an incentive design, because that enabled just a minority of people in the large audience to also remind other people to take care of themselves,

essentially appealing to their self-interest by protecting themselves against their own hunts. And based on water usage data, we know that it has worked because people have washed their hands much more vigorously than before across all the urban and rural areas. And I think all this is communicated very clearly by our Central Epidemic Monson, the CECC, which hosts a daily press conference ever since January

for 140 days, which is always live streamed. They answer all the questions from a journalist community, and there's a hotline, one night to do, where everybody can call them with new social innovations. For example, there was one day in April where a young boy who said he doesn't want to go to school because his schoolmates may laugh at him because when you ration the mask, you don't get to pick the color,

and all he had is pink medical mask. And so the very next day, everybody in the CECC press conference are wearing pink medical mask, making sure that everybody learn about gender mainstreaming, which is also a social innovation and is on the basis of this trust that the people who built, for example, a interactive map that showed mask availability in the nearby convenience stores, even though they only managed to pay for that website on the GovZero Slack channel. There's many people working on it, but because they forgot on the Google Maps,

place API, they can cash the place ID. They end up owing Google a large sum of money because tens of thousands of people started using it. And so I brought that map to the Premier who says, oh, we need to support this civic innovator because they have a better idea than our own. And so we work with our National Health Insurance Agency who has a card, a ICOT that covers more than 99.99% of people in not just citizens, but also residents, so that they can go to their nearby pharmacy, swipe their NHI card, get nowadays nine masks per two week, if they're an adult and 10, if they're a child, medical mask. And we make sure that the stock level of all the pharmacy for adult and children masks are published every 30 seconds in the beginning,

now it's every three minutes. So it's almost like the distributed ledgers, which enabled more than 130 tools, not just masks, but also voice assistant for people with blindness, who can get the same inclusive access to an information about which pharmacy near them still have medical mask. And because of that, people who have any symptom will be safely taking a medical mask, going to a local clinic, knowing surely that the single-payer system will take care of them, whether they actually are hit with coronavirus or just some other disease. Just curious how that works. In the United States, there's private pharmacies like CVS or Walgreens or things like that. I'm just thinking, you know, CBS and Walgreens don't have a data feed of how many

masks are available. I'm assuming the pharmacies in Taiwan are private companies as well? How does that get to be made hackable by the civic tech hackers who again developed a solution on the outside and was brought back to the inside of government? We know from the post office records which pharmacy receive how many masks in supply early in the morning and because of that we have accurate stock level information for the different medical masks. So when the person make a purchase, everybody standing in the can check that actually it has decreased by nine after 30 seconds. And this is participatory accountability. If they see rather the stock level of the pharmacy rises, they will then call 1922 reporting something is wrong. So the fairness is

guaranteed not just by the words of the Minister of Health and Welfare, but actually by everybody who can participate in the distributed ledger and keeping each other accountable. So Audrey, there's something there that I think really speaks to our present moment that I'd love to hear from you about. You've described yourself often as a conservative anarchist, which is an idea that resonates with me quite a bit. And one thing that's always struck me about Taiwan is both the low number of police and the low amount of crime. And obviously we're in a moment in the United States where there is a pushback against policing. It's always struck me that what was critical to achieving that in Taiwan was precisely

the sort of overlapping community monitoring that constitutes for police presence. I wonder whether you see a connection between what you just described in terms of sort of economic allocations and law enforcement. Yeah, definitely. I think in Taiwan, the norm is pretty strong in the sense that every person who see any unfairness on the street, even just as simple as a car parking out of its parking space, they will take a photo and basically do a social sanction stuff, which may or may not have its drawbacks, especially if they complain about the wrong things. But by and large, I think people correct themselves when they found that it's actually not what it seems. And so it's in general a positive saying that people will determine the norm regulation basically together and investigate and resolve. Essentially, civilian complaints around law enforcement, whether there's a over-enforcement or under-enforcement by, for example, the community, public servants in the district level,

who help with, for example, home quarantine people. People who are in the home quarantine for 14 days, they can choose to go to a hotel, of course. But if they prefer to stay at home, they can put their phone essentially into the digital fence. And if their phone breaks out of the 50-meter radius, which is determined not by GPS or an app, but rather by cell phone tower triangulation,

then the local public servants receive that SMS from the telecoms and will then go physically within minutes to check the whereabouts. and check the health and so on. And that instead of a very heavy-hunted penalty or criminal offense, they will often just take care of the mental health, the mental needs of the person in home quarantine so they can stay the full 14 days. There's many examples by home quarantine. It's a really good example. Just to double underline something in your password is just how responsive the systems are to problems. I mean, you know, how quickly you're getting feedback.

I think just by contrast in so many other countries, you don't have that level of speed feedback loop between government implementing a solution, between people proposing and saying there's a problem here, whether it's an Uber policy or air pollution. Do you want to speak briefly to? Certainly, certainly, because when people talk about public-private partnership or sometimes public-private partnership, right, that's also a thing. I think that the order is quite wrong. So in traditional public-private partnership, the public sector came up with an idea and the private sector implements the idea and maybe the people, that's to say the social sector, work with making sure that it integrates well with the local norm and things like that. But that's by nature slow,

just as with any procurement process. But in Taiwan, this is the reverse. Essentially, the social sector builds a prototype such as the musk nut, even though they may not have the resources to maintain it for very long, but they can work with the activists and the media to make sure that everybody in Taiwan learns is a good idea. And then the private sector can step in, saying, for example, Google said, okay, we waive all the way API usage because it's for public good. And that also put pressure on the public sector for making the open data possible because now even Google promised to say that we will host essentially the mask map for free. And so the public sector, all we have to do is essentially on the side of like a vendor,

right? The specification is already written by the social sector. There's already rough consensus. We already know that people want the mask map. So all we have to do is to short the interval on which we published the pharmacy stock level information from one day to 30 seconds. That's literally the only thing we did. And so because of that, that happened in 48 hours. And part of that is having a digital minister and a team like yours to be able to implement something like that. Is that correct? Well, we work, of course, with still contractors and things like that. But basically because the blueprint is already provided by the Go-Zero community, essentially the forked government is already there and because you always relinquished the copyright associated with that. So it basically

just took their code and did the maintenance part of the backend infrastructure, which doesn't take a lot of lines of code. Yeah, I mean, I think what's really remarkable about this, and again, this goes to the name conservative anarchism, is that if you have effective tools for building and signaling consensus, actually a lot of the tools of formal power come to sort of seem less important. And that sounds a bit strange, but actually historically, it's very consistent with the way the democracy evolved. The parliament in Britain arose out of the king's attempt to gain information about what was going on in the kingdom. And only as it became increasingly relied upon did it then become a formal power structure. And I think we're seeing something

similar where on the one hand, it might seem like you just are inventing some tools to help people get more information, but on the other hand, it's sort of becoming its own system of government. Yeah, this is what we call listening at scale. And to the MPs, I always say that in design thinking terms, this solves the problem of how might we go forward, which is about defining the issues. But of course, the delivering of the issues and the development of the laws and regulations is still the MP's purview. It's just like the MP can now drive with more visibility where everybody can see where we're going as a policy and people coming to a consensus can inform the driver much more easily. And eventually, of course, the goal for me personally is

make it a self-driving vehicle, meaning that we did away with the entire political class altogether, but I realized that may or may not happen in my lifetime. For now, we coexist peacefully with representative democracy. It's sort of like it's hard to drive a car if the two people in the backseat giving you directions keep disagreeing about where they want to go. Exactly. Exactly. So just give them GPS, you know, map software. It reminds me of Larry Lessig's work that, you know, he has a book called, they don't represent us. But his second point is that we don't represent us, that we actually agree more than we think we do. But none of the attention and the ways that our speech are represented in the attention economy on the digital tools that we use, show or select

where we actually have agreement. And like you said, how can you, as a government official, argue with a majority of huge consensus about what needs to be done? As soon as that consensus is visible, so if we have tools that make, you know, superhuman consensus levels, you know, available, then that solves so many other problems downstream. And suddenly you don't have people arguing in the backseat. You have exact consensus about where you're wanting to go.

Yeah, exactly. And that metaphor, by the way, except for the self-driving carpet, is from Dr. Siasen, so founder of this country. So one of the other fascinating things about Taiwan, given that it relies on, you know, a similar open online environment for deliberative systems and the fact that it's one of the biggest targets for China's disinformation and propaganda campaign, you would think that whether it's in the coronavirus response or in any of the major sort of policies that it's proposing or in the election, it would be, you know, even more bad disinformation just flowing through the system more so than even the United States. How has it given that seemingly weathered the storm? You have a bunch of powerful ideas about how you were able to mitigate some of that. Yes. So essentially we look at the infidemic like a epidemic.

So there's roughly speaking two sources. One is domestic, which often stems from outrage, which is one of the most potent emotions. If people can feel that they really need to share this outrageous situation with other people, they would not bother to fact-check the content, and then it will very quickly go viral, quite literally, and inciting even more outrage and more divisiveness and polarization. And so that's one side, and the other side, of course, is full-fledged propaganda from the CCP, which I'll get to later. And so for the first part, which is just outrage, I'll use one concrete example to illustrate our way of countering it,

which is called humor over rumor. So there was a rumor that says the tissue paper material are being confiscated, nationalized by the state to make medical masks because we're ramping up production from 2 million medical masks a day to 20 million, we will soon run out of medical mask material. And because of that tissue paper will run out even sooner than that. And because of that, you need to rush out and buy tissue papers, unquote. And so it caused a lot of outrage and stress and panic, right? and conspiracy theories.

And so we have this triple two principle that says every time we detect such a rumor, maybe people call Y922 asking whether if it's true, within two hours we need to respond with two pictures that each have two hundred characters or less that are funny, that makes people laugh. And so our premier, Suzen Chang, who wiggles his bottom in a meme, gets rolled out within a couple of hours and says in very large print. that we only have one pair of Botox each, meaning that we don't need to panic by tissue papers.

And coupled with that a very clear table that says facial mask are produced using domestic material and tissue paper using South American materials. And that went absolutely viral, not only because of its weakling button, but because it's a brilliant pun because bottom and stockpiling is a homonym. It sounds the same in Mandarin. So anyway, it's very funny, professional comedians. So in any case, that went absolutely viral. And because of that, everybody who laughed about it is literally unable to feel outrage

when they see the conspiracy theory later because the pathways in the brain, they cannot coexist if you have watched the film inside out, you see what I'm talking about. So the peggy buying of tissue papers die down within a day or two. And we found out the person who spreads a rumor at the first place was a tissue paper reseller. We make sure our humor has a higher R value than the conspiracy theory. theory is naturally vaccinating the society about that. And so that's the fun part. Now, there's, of course, the less fun part, which are just deliberate campaigns of narratives. For example, last November, right before our presidential election, there was a really viral message on

the Taiwanese social media that says, the rioters in Hong Kong are paying young people 20 million to murder police. And so what we did is not making fun of it. There's nothing fun about it. What we did is working with international fact-checking network, the IFCN journalists, the Taiwan Tech Check Center, the TFCC, did an attribution work to find out who was the original person who repurposed this neutral Reuters photo, which says nothing in its caption about paying to murder police into such a weaponized information. And so they discovered that originated the very first post as from the social media of the essentially propaganda units of the Chinese Communist Party.

And because of that, we work with platform economies, for example, Facebook, to make sure that everybody who see this part of mold information see a little note behind that photo that says the TFCC have checked, and this mold information is being sponsored by the Chinese Communist Party. And so we did not do take down, but rather we make sure that people learn that this is essentially political propaganda. And the Reuters, of course, says that, you know, the caption is not from them. And so I'm sure that people who see this, they still share this outrage, but outrage will be framed very differently than if they only see the misleading caption. Just curious, did they back notify everybody who had also previously seen that information, so not just future looking? That's right. So it retroactively adds this fact-checking notice on the social media. Just as during the presidential election for the first time, Facebook agreed to publish according to our social norm. all the political advertisement and social advertisement during the campaign session, including

who they micro-targets, who they work with, for example, who sponsor it, and also, of course, it needs to come only from domestic sources, just like our campaign financing, and most importantly, how many people did they actually reach? And with this real-time information, the Gov0 people who work on the voting guide, make sure that anybody who would try the dark patterns of, you know, micro-targeting some people and fuel them whose conspiracy theories would get named and shamed. And because of that, that did not happen during our presidential election. Well, there's so many fascinating things that you brought up in that whole segment. On the first side of sort of organic misinformation, when you take this sort of info-epidemiology view,

that outrage has a higher R-Not that spreads virally than non-outrage. The only way to combat that is with sweetening on a different size, the R-not of with humor. And in comedy, it reminds me of the work of, of Sierda Popovic, who's the, I don't know if you know his work, he's the founder of laftivism and using laughter to go against authoritarian dictators who will otherwise clamp down on protesters. So the second thing I love about your humor strategy is, you know, about a misinformation or a rumor or a conspiracy, is you mentioned that you don't have the humor make fun of the person who's spreading the conspiracy theory. Is that right?

That's right. So the premier makes himself the butt of the joke, so to speak. And of course, quite literally the butt of that joke. Right, right. That's right. That's right. So it's humor. It's not satire or parody, right? It's humor as it's originally meant. But what it struck me is that if you have to respond to rumors within two hours, you're still gated by, well, how fast can I generate bad viral rumors? Because if you have a bad actor that can outcompete your team of comedy writers, you know, what do you do in that circumstance? Well, then, of course, if this is organic, then we just send an invitation letter to the people who protest and say, why wouldn't you like to co-create our policy? That always work. So, for example, for people who complain that the text filing experience was explosively useful, they actually may have a point. And so their outrage have a point. So we don't mock them. We rather invite everybody who complained and say, let's co-create a tax filing system for the next year.

And that always work if it's from a domestic or organic source. What doesn't work, of course, is from, you know, people who refuse to participate in the democratic process, sometimes from overseas. And for that, of course, laughter remains a strong strategy. And we make sure that the premier, many ministers, including myself, pretty clear, their likes their images for memetic purposes so that the comedians don't have to check our copyright or whatever. Wow.

It's so fascinating what you're saying. I mean, can you just imagine if for every time there's genuine outrage that you simply can't follow with humor, you say, hey, I'm actually going to invite you with that outrage into saying, how do we change the system to actually meet those needs? And you actually, this is what you do, right? You invite them into a group in which they start co-creating an actual policy to solve. That's their problem. That's right. And that works in a country of 23 million people. So you have, you know, sort of scales of outrage and a number of topics that somehow managed to fit within the scale.

Yes, definitely. And the more, the more outrage. there was in the first place, the more people participate. So our tax filing system, which was a case that we invited to co-creation around three years ago, eventually built the new taxonomy experience two years ago that have like 94% of approval rating, which is unheard of in digital services rolled out by the government, precisely because thousands of people can say, hey, I contribute to that, at least one post-it note,

was from our co-creation workshop or from my slide of comments or from my live-stream comments. comments. And that's my idea. And because of that, they voluntarily taught their friends and families of how to make the use of the tax violence system that they co-designed. That's amazing. So you have both built yourself as a hacker and as a political advocate and as a movement leader. How would you, seeing everything that's going on in the West with private technology companies, social media, amplifying disinformation, and in the way that governments are not really structured to do the things that you're doing? What lessons would you be applying to both our private technology companies and to the way that, you know,

we might be thinking about fixing our government right now? Yeah, so I helped recording a series of lectures, a set of strategies for a 21st century lawmaking. And the basic idea is that technologies can engage the public to improve the quality and legitimacy of lawmaking, even as simple as just, you know, troll control, right, dealing with the astrotruvers and things like that. take away the reply button. So very simple mechanisms that once applied to just remove this whole set of issues that makes the lawmakers think that public engagement is such a chore,

it's not time consuming, and so on. So my main lesson to teach on that is just that there are always time-saving mechanisms. The other thing is that it's more legitimacy. So if you work with the populism, which I think the problem is not that it's popular, is that they tend to exclude people from the population, but if inclusive populism, like including people under 18, which are very active in our participation platforms, people even who are not born yet, and people who advocate for them, people who are people, but traditionally seen as mountains and rivers, because we have an indigenous tradition that look at the Savilla, the highest mountain and see them as essentially people, natural personhood, and so on. If you include these into more inclusive

populism, then you can get a much higher legitimacy because there's kind of by definition, just as Glenn mentioned, a supermajority of trust that supports you in your decision-making. And this has been shown to work in the U.S. in Bowling Green Kennedy. They run a civic town hall, a virtual town hall using the same police technology, and they revealed, for example, no matter people identify as Republican or Democrat, everybody agreed that you should include arts in the science, technology, engineering, and mass. education. Now you don't see that on the half-wise, but that's something that actually both I care a lot about. So if the mayor just do that, they kind of automatically gains

more trust and there's less political risk. And so more trust, less risk, save more time. These three are the three pillars of the Pareto improvement, making that improvement on one of the three axes without sacrificing, the other two, and it can always be done in a piecemeal way, just as we did, post-occupying. And how would, government's, imagine there is cities and states that wanted to start implementing this. I mean, how would they actually implement something like V-Taiwan and pull that is? Yeah, you can read the Crowell Handbook. Yeah, and Radical Exchange as well.

So GovLab is one organization and Radical Exchange is another. We've been working with the government in Colorado at many different levels. So there are several groups in the United States that are working to build the census around this. And I think they complement each other very well. I mean, if you're from a parliament restructuring viewpoint, like you want to improve the structure of how representative democracies function. The set of tools is basically tried and true. If you are on a smaller scale, like if you're a township, then the radical exchange ideas are much more economically efficient, but have not been as rigorously tested. But it will be great because you will be making academic contributions.

Wonderful. Audrey, thank you so much for joining us on the podcast. It was wonderful to have you. Thank you, Glenn, so much for assisting in being here. Okay. Bye. Bye. Okay.

Imagine that civic hackers actually took the tools that Audrey's talking about and build parallel websites to our existing city and government websites here in the United States, civic tools that were actually built for the public interest. That's what Audrey's work is about, showing that it's actually possible to have technology and democracy coexist in a way that does not lead to the kind of dystopia that we tend to talk about here on your undivided attention. There is so much to learn from her example. And one of my biggest hopes for this episode is that people really take it to heart and imagine what it would look like for every city council building and every state building and every mayor and every senator and

every state senator to start using and applying these tools to actually show that democratic governments can actually work in the 21st century in a digital context. And if you'd like to to learn more, we really highly recommend that you check out Audrey's work. Audrey has the white papers online on our website talking exactly about how they implemented each of the tools and how it worked. Your undivided attention is produced by the Center for Humane Technology. Our executive producer is Dan Kedmi and our associate. producer Natalie Jones, Nor Al Samurai helped with a fact-checking, original music and sound

design by Ryan and Hayes Holiday, and a special thanks to the whole Center for Humane Technology team for making this podcast possible. A very special thanks to the generous lead supporters of our work at the Center for Humane Technology, including the Omidiar Network, the Gerald Schwartz and Heather Reisman Foundation, the Patrick J. McGovern Foundation, Evolve Foundation, Craig Newmark Philanthropies, and Knight Foundation, among many others. Huge thanks from all of us.