#163 - Interview
🇬🇧 How Taiwan reinvented its democracy
Technology, Collective Intelligence and Political Resilience with Audrey Tang, Former Taiwanese Minister
Écouter sur :
🇫🇷 fr
NB : la VF est disponible si vous ne comprenez pas anglais.
Audrey Tang is Taiwan's former Minister of Digital Affairs and now serves as its cyber ambassador. A self-taught hacker and Taoist thinker, she found herself at the heart of a unique democratic experiment in a country under constant geopolitical pressure.
Why do our democracies seem so fragile today, when we have never had so much information and so many communication tools at our disposal?
Mistrust, polarization, a sense of collective powerlessness.
Beyond parties and leaders, it is perhaps our way of producing meaning and making decisions together that is faltering, in an ecosystem dominated by private platforms and now by artificial intelligence.
In Taiwan, another path has been explored: viewing democracy as a living system, capable of learning, listening, and evolving, even under pressure from disinformation and cyberattacks.
In this episode, we talk about technology, emotions, collective intelligence, and trust.
And what democracy could become if we fundamentally rethink the way we connect with each other.
Episode 1
0:32 Fragile Democracies
1:43 Mass Extinction Awareness
2:19 Audrey Tang's Journey
6:42 Legacy and Power Sharing
13:59 Examining Democratic Struggles
20:00 Purpose of Democracy
22:24 The Role of Representation
27:03 Collective Intelligence in Action
29:38 The Sunflower Movement
36:03 Digital Tools for Democracy
40:05 Bridging Polarization
46:18 Trust and Governance
49:14 The Living Democracy
53:13 Scaling New Protocols
57:29 Redemption in Social Media
Episode 2
0:03 Introduction to the Conversation
3:47 Emotions in Politics
5:44 The Role of Technology
8:23 AI and Assistive Intelligence
14:31 Radical Transparency in Democracy
17:04 Education and Civic Engagement
20:13 Reshaping Education in Taiwan
20:49 Lessons from Taiwan
23:20 Technology and Governance
25:06 Competing Models of Control
30:55 Tensions in AI Usage
34:22 Conversations on AI Governance
40:31 Navigating Corporate Agendas
46:19 Six-Pack of Care Principles
52:44 Protecting Relational Health
54:12 Reinventing Public Spaces
55:10 Imagination and Hope
57:00 Parenting and Technology
59:54 Meaning of Life and Legacy
1:02:13 Recommended Readings
1. Democracy as a Living System
Democracy is not only an institutional regime, it is a collective learning process.
Its role is not merely to distribute power, but to generate shared meaning in a complex world.
A healthy democracy increases its “bandwidth” for listening and reduces its “latency” in responding to social signals.
2. The Structural Problem of the Information Ecosystem
The attention economy fragments public debate by rewarding outrage rather than deliberation.
Private platforms have become central mediators between citizens, information, and power.
The issue is not only ideological, it is architectural, rooted in system design.
3. Emotions and Collective Intelligence
Emotions are not noise to eliminate, but data to integrate into collective decision-making.
Designing protocols that make areas of convergence visible helps move beyond polarization.
Collective intelligence emerges when disagreements are structured rather than amplified.
4. The Taiwanese Experience
In Taiwan, tools such as vTaiwan and Pol.is enabled large-scale discussions without descending into chaos.
Trust is a matter of design, it is built through transparent and open protocols.
The state can become a facilitator of cooperation rather than merely a command center.
5. Openness and Security
Defending democracy against disinformation does not require closing the public sphere.
Resilience depends on civic competence, not only on censorship or surveillance.
Civic and technological education is a long-term strategic pillar.
6. AI, Power, and Decentralization
AI can expand participation or reinforce manipulation, depending on design principles.
The tension between profit, attention capture, and democratic interest is central.
Imagining a pluralistic AI means distributing intelligence rather than concentrating it.
7. The Challenge of Technopolitical Models
Between state control and platform capitalism, a democratic space remains possible.
Europe must move beyond regulation to invent its own civic infrastructures.
Concepts
Deliberative Democracy
A model of democracy that prioritizes structured discussion among citizens before political decisions are made.
Collective Intelligence
The capacity of a group to generate better decisions or insights than isolated individuals, when the right framework and protocols are in place.
Attention Economy
The dominant model of digital platforms, based on capturing, monetizing, and competing for users’ attention.
Plurality / Plural Technology
A concept developed by Audrey Tang and E. Glen Weyl aimed at moving beyond binary thinking and enabling multiple perspectives to coexist within digital systems.
Book:
Plurality: The Future of Collaborative Technology and Democracy
Publisher (Princeton University Press):
https://press.princeton.edu/books/hardcover/9780691245935/plurality
People
Audrey Tang
Former Digital Minister of Taiwan, now Taiwan’s Cyber Ambassador and an international figure in civic technology and digital democracy.
E. Glen Weyl
Economist, co-founder of RadicalxChange, and co-author of Plurality.
https://www.radicalxchange.org/
Sam Altman
CEO of OpenAI.
Gavin Newsom
Governor of California, engaged in AI regulation and innovation policy.
Tsai Ing-wen (蔡英文)
President of Taiwan from 2016 to 2024. She supported Taiwan’s open digital governance model and collaboration with civil society.
https://english.president.gov.tw/
Organizations and Tools
g0v (Gov Zero)
A Taiwanese civic tech community developing open-source tools to increase government transparency and public participation.
vTaiwan
A public consultation platform that enables citizens and policymakers to co-create legislation through structured online deliberation.
A large-scale deliberation tool designed to identify areas of consensus within polarized debates.
Join Platform (Taiwan)
The Taiwanese government’s official citizen participation platform.
PDIS – Public Digital Innovation Space
A government innovation unit established under Audrey Tang to promote open digital governance.
Cofacts
A collaborative fact-checking platform used in Taiwan to counter disinformation.
RadicalxChange Foundation
An organization working on democratic governance and new economic architectures.
https://www.radicalxchange.org/
Historical and Geopolitical Context
Sunflower Movement (2014)
A student-led occupation of Taiwan’s Parliament protesting a trade agreement with China. It became a foundational moment for Taiwan’s civic tech ecosystem.
Chinese Pressure and Disinformation
Taiwan faces ongoing cyberattacks and influence operations from China, shaping its approach to democratic resilience.
Digital Services Act (DSA) – European Union
EU regulation governing platform responsibility and digital accountability.
https://digital-strategy.ec.europa.eu/en/policies/digital-services-act-package
AI Act – European Union
The European regulatory framework for artificial intelligence.
https://digital-strategy.ec.europa.eu/en/policies/european-approach-artificial-intelligence
1. Dao De Jing (Tao Te Ching) – Lao Zi
Audrey Tang specifically recommends the English translation by Ursula K. Le Guin, which she describes as “pure poetry.”
A foundational text of Taoism, the Dao De Jing explores the flow of the Dao, the nature of reality, and the principle of wu wei, often translated as “effortless action.”
Babelio (Tao Te King):
https://www.babelio.com/livres/Lao-Tseu-Tao-Te-King/1912
Ursula K. Le Guin translation (Shambhala Publications):
https://www.shambhala.com/tao-te-ching-9781611807240.html
2. Yijing (I Ching) – The Book of Changes
Also known as The Book of Changes.
Audrey describes it as an ancient “binary code,” a symbolic matrix for understanding relationships and transformations within reality.
It is not meant to be read linearly from beginning to end, but rather consulted as a system, a navigational tool through patterns and change.
Babelio (Yi King / I Ching):
https://www.babelio.com/livres/Anonyme-Le-Yi-King--Livre-des-transformations/19746
Audrey Tang and Her Work
Audrey Tang – Official Website
Speeches, essays, interviews, and initiatives related to digital democracy, civic technology, and pluralism.
Plurality – E. Glen Weyl & Audrey Tang
A proposal to redesign collaborative technologies beyond binary systems, enabling pluralistic digital infrastructures.
Princeton University Press:
https://press.princeton.edu/books/hardcover/9780691245935/plurality
Hacking, Open Source, and Governance
Steven Levy – Hackers: Heroes of the Computer Revolution
A foundational book on early hacker culture and the ethics of openness and information freedom.
Penguin Random House:
https://www.penguinrandomhouse.com/books/304420/hackers-by-steven-levy/
Yochai Benkler – The Wealth of Networks
A landmark analysis of networked collaboration and peer production in the digital age.
Open-access edition (Harvard):
https://cyber.harvard.edu/wealth_of_networks/
Lawrence Lessig – Code and Other Laws of Cyberspace
Explores how code regulates behavior as powerfully as law, shaping digital governance.
Basic Books:
https://www.basicbooks.com/titles/lawrence-lessig/code/9780465039142/
Democracy and Deliberation
Jürgen Habermas – The Structural Transformation of the Public Sphere
A foundational work on the concept of the public sphere and deliberative democracy.
MIT Press:
https://mitpress.mit.edu/9780262581080/the-structural-transformation-of-the-public-sphere/
Hélène Landemore – Open Democracy
A defense of collective intelligence and large-scale participatory processes.
Princeton University Press:
https://press.princeton.edu/books/hardcover/9780691181998/open-democracy
James Fishkin – Democracy When the People Are Thinking
Research on deliberative polling and structured citizen participation.
Stanford University Press:
https://www.sup.org/books/title/?id=26364
Education and Digital Culture
Howard Rheingold – Net Smart
On developing critical and collaborative digital literacy.
MIT Press:
https://mitpress.mit.edu/9780262529839/net-smart/
Danah Boyd – It’s Complicated: The Social Lives of Networked Teens
A sociological exploration of youth and social media culture.
Yale University Press:
https://yalebooks.yale.edu/book/9780300166316/its-complicated/
Taoism and Non-Dual Thinking
Lao Zi – Tao Te Ching
Foundational Taoist text exploring harmony, flow, and non-coercive action.
Shambhala (Ursula K. Le Guin translation):
https://www.shambhala.com/tao-te-ching-9781611807240.html
Zhuangzi – The Complete Writings
A deeper philosophical exploration of Taoist thought and perspective fluidity.
Hackett Publishing (Burton Watson translation):
https://hackettpublishing.com/the-complete-works-of-chuang-tzu
Collective Intelligence and Polarization
Cass Sunstein – #Republic
On information fragmentation and echo chambers in the digital age.
Princeton University Press:
https://press.princeton.edu/books/paperback/9780691175515/republic
Eli Pariser – The Filter Bubble
On algorithmic filtering and personalization online.
Penguin Random House:
https://www.penguinrandomhouse.com/books/307207/the-filter-bubble-by-eli-pariser/
Julien:
Hello, Audrey.
Audrey Tang:
Hello. Good local time. Very good to be here.
Julien:
Yes. So you're in Barcelona right now, in Lisbon. We could almost have been together in the same place, even though it's a few hours to drive to you. So we're doing this online, but normally not often here. You don't live here. You're from Taiwan, right?
Audrey Tang:
Well, I'm originally from Taiwan, and I served as digital minister in the cabinet for seven and a half years. But now I am cyber ambassador at large, which means I traveled to 27 countries in the past year, and I switch time zones every five days.
Julien:
Oh, wow. Okay. So difficult to define where is home to you, I guess, now.
Audrey Tang:
Planetary.
Julien:
I love it. So, I mean, your life is quite amazing and even extraordinary because you left school.
Julien:
I would spoil a little bit, but then you can talk over that. But you left school as a teenager you learned code almost before you could read uh you became a hacker you know obviously and then a minister and somehow you managed to stay even a poet through all that yes i want to to start talking about you yeah let's about you and your your ideas before going into your work and then we spent a lot of time talking about what you did and what you're doing now but my usual question is about your lenses you know how you look at things so when you look at the world today, what are these lenses like, what do you use to make sense of it.
Audrey Tang:
Well, I see the world as a network of living relationships. Everything from an ecosystem to a city, from a piece of code to community, is a set of relationships that has relational health. So it's like a Tao, a Wei, a flow that is constantly exchanging. So my lens, I would say, is relational. I do not focus on the individual themselves, but rather on the connections between the various individuals.
Julien:
Okay, so you try to spot the links in everything and the patterns, I guess.
Audrey Tang:
Yes, it's a very Taoist way of thinking. And as you alluded to, I've been an autodidact since my teenage days, and I've been practicing Taoism as a survival skill since when I was four.
Julien:
Yeah, we can go into this because, of course, that has shaped you, I guess. And you have a high... I want to understand how these very early experiences, so the illness, and you can say a few words about this, meditation practice and your Taoist also upbringing, how this shaped the way you think about power, technology, and corporations and everything in general.
Audrey Tang:
Hmm. Yeah, to me, my childhood illness, the early diagnosis when I was four, saying that this child only have a 50% chance to survive until heart surgery. I think that's really defining because ever since then, every time I go to sleep, even now, I feel like flipping a coin. If it doesn't land well, then I don't wake up the next day. And so I got into this habit of publishing before I perish. So I record everything I learned every day, first on cassette tapes, floppy disks, and finally the internet. And so we've just been talking about before entering the recorded part, how all the transcripts for the past 10 years is online in the public domain, over 2,000 conversations with 8,000 people. And I took it almost as a daily ritual to ensure that I publish these before I go to sleep. So even if I don't wake up, that's fine because that's a different form of power. It is a power of a good enough ancestor that leaves the next generation with more material, more canvas, but it's not powering over them to control how they use these materials.
Julien:
So since you were very young, you have this in mind. You have this worry, which is about not waking up the next day. And, you know, what did you learn? How did they shape you in the way you go into a conversation, you go into a problem? You're talking about, thinking about what will be your heritage, your legacy, but what else is it shaping that sets you apart, from what you understand from the others.
Audrey Tang:
Yeah, I wouldn't say that it's legacy per se. Rather, it is a commitment to sharing power because accumulation makes no sense if you're not going to wake up the next day, right? So, and the Taoist breathing exercise is to keep me survivable until I got a surgery, which I did when I was 12, because if my heart beat above a certain beats per minute when I was a child, we'd just faint and wake up in the hospital or something. And so I had to be extremely homeostatic in the sense of keeping to the energetic middle ground, not too joyful, but also not too angry, not controlling my new details, but also not out of control and being taken by passion.
Audrey Tang:
So like everything has to be kind of in the middle and moderate for me. So that is the Taoist idea of wu wei or effortless action, not forcing a solution, but also not detaching from conflict. It's about listening deeply about the entire system so you understand its natural flow.
Julien:
And this is something that you were able to describe very early in your life, like that approach to things?
Audrey Tang:
Yes, I was quite articulate in that. And I think it is really helped by the fact that I was very much into math,
Audrey Tang:
Mathematics, and also into coding. So to me, a personal computer, which was born the same year I was born, is kind of like a musical instrument that I can take as its notes,
Audrey Tang:
The computer program, but then the output, like a symphony, is the way people interact. And very quickly after programming quite a few games where other kids can use to learn mathematical concepts and interact with each other, I found that there really are like pro-social ways to arrange interactions in a game and also anti-social ways to design a game so people feel the need to compete in a negative sum or a zero sum way. And so I think the articulation from my side was helped by this instrument that is both very precise in a sense of mechanical, like almost engineering mindset, but also it is very creative in a sense of the human relationships it can foster. Of course, that then later led to the internet. And so I started a company when I was 15 with a couple of other friends. I was the chief technology officer, and we were the first firm in Taiwan to do online people-to-people auction like eBay. We call it Qubit, an instant messaging and search engine and things like that. Again, tapping on the ability for internet to connect people so people swiftly trust each other in that pro-social way. as kind of enjoying the same space together.
Julien:
So very early, you started to meditate, then you started to code, and very early, you had that preoccupation about connections. I want you to understand where this comes from, because a lot of people start to code and make very different things with it. But from what you're telling me, very early on, it was about how can I make people connect in a better way, right?
Audrey Tang:
Yes. And also, that was because, as I mentioned, I practice meditation, not because it's a hobby or something, but rather as a revival skill. So I had to kind of run this internal inner system loop to be very attentive, not just to the environmental stimulus, but also to my inner state, the loops of anger, the loops of outrage, the loop of fight or flight, and debugging them, refactoring my consciousness for clarity and always get a long night's sleep. At least eight hours, so that these learnings get written into the long-term memory through neuroplasticity, right? So to me, meditation is like coding. It requires a very attentive way to look at the entire system and also to come up with a set of clear metaphors or simple rules that allows for complex emergent behavior.
Julien:
So this is what defines kind of your worldview today. You talked about connections. You talked about seeing things as systems, understanding the rules. So that's what you're passionate about. I understand you understand how you try to look at the matrix kind of. But again, why social things became so central to you? why social interactions became one of the key topics that led you to be where you are today.
Audrey Tang:
So I think the relational ideas in a Taoist thinking is looking at the conflicts, the contrasts, the differences, the yin and the yang, not as fire to be put out or to stay away from, not as volcanic explosions, but rather as energy, as like magma underneath. So if you do meditation well, if you do your breathing exercise well, then you become heat resistant enough so you can become a channel that these conflicts flow through you. And then almost like a geothermal engine turns that heat into the energy of co-creation from the yin and yang comes the 10,000 things, right? So that is the main metaphor. So instead of calling it social, which sounds like, I don't know, fitting into an existing script, I prefer to call it plural, as in plurality, as in collaboration across differences and as in upholding the space in which conflict is seen as energy instead of as bugs. Conflicts are features, not bugs.
Julien:
Okay. Well, I mean, through the conversation, feel free to explain where the ideas come from, because I think that's very interesting to refer to your context, your starting points. And I like to have that, not just ideas, but understanding where they come from and what are the potential conflicts that you're going through when you think about the whole thing. I want to talk about democracies, and we can spend a lot of time on this, because obviously that's one of the core topics that keeps you busy at night. So everywhere we look now, we can see that almost all democracies are kind of struggling. They seem to be struggling, because we talk about polarization, there is mistrust, and people seem to be kind of exhausted. You know, like the system is being exhausted, tested. How do you see that moment? You know, what does this moment in time reveal to you about the potential deeper limits of our current systems?
Audrey Tang:
Well, the exhaustion I believe is a symptom of a mismatch between the old operating system of our democracies and the current system that we're heading into. The old system, the one that people was used to, were broadcasting systems in an age of scarcity of information. A few people, gatekeepers, talk and everybody listens. And the feedback loop is measured in a number of years through voting. But now we live in a very different era because it is a networked communication with information abundance. For most people, the polarization, the friction that you see is just a symptom of people trying to fit the old system into this loop of more and more information being generated beyond what we can digest. And 10 years ago, people came up in social media, this idea of a recommendation system. So instead of people subscribing, following the people who push out content, so if we subscribe to the same person, we see the same world.
Audrey Tang:
This time, the AI system figured out how to keep us engaged to our screens. And more often than not, they keep us enraged, engagement through enragement, by amplifying what I call PPM or polarization per minute so that people experience a very high polarization per minute because that's the thing that keeps people glued to the screen. So instead of a few gatekeepers making sense of the daily news, now people are human in the loop of AI, like a hamster in a hamster wheel. And so the hamster wheel turns faster and faster. The hamster feels great because probably needs to exercise, but has zero control, zero steering of where the hamster wheel is going. And so the system is struggling because the old system, this broadcasting system, applied to this massive information abundance, just prioritizes the megaphone to the extremes. Because this system was never designed for broad listening, for listening at scale. It was just doing broadcasting at scale. So everybody broadcasts, retweets, dunks, but nobody really has the time and space to pause and listen broadly in this space. you
Julien:
So it's a mix between the attention economy that basically steals all the attention that people need to be able to listen, to reflect on things. And it's also a revolution related to the fact that the gatekeepers are no longer setting up the agenda. Right um well before going diving into this a big question is related to democracy itself what's what do you see being the purpose of democracy and what is what is it ultimately you know meant to achieve for for society you know is it what are we trying to do here in the end. It's not just about maintaining democracy, it's about maintaining something, making something work.
Audrey Tang:
Yes. To me, democracy is like a society's nervous system. And the job, of course, is both to sense the needs, the pain, the wisdom of all the different parts, especially the most vulnerable. But then it's also not just sensing, but also sense-making, processing that information into a coherent understanding so that it creates common knowledge in the society. So now I know that you know, that I know, that everybody knows that this is happening, so that we can take coordinated action on it. So the ultimate goal is not just winning one election or something. It is to make sure that people understand that we, the people, are already the superintelligence we've been waiting for. So we don't need a superintelligence robot to solve all our problems if we have good enough relational health, if we're not masked by the very high polarization per minute, a high PPM, then we can actually get together in a way that's fast, that's fair, that's fun, and turn the heat, again, of the social conflict into sensing and then sense-making.
Julien:
So in the end, the idea is to find a system that makes it possible to take the best decision for the group, to maintain society together. And I like what you said about the fact that the current operating system basically is being challenged because of technology in part. And we can say a few words about that system because modern democracies were never really designed to give direct power to citizens. Because in the early forms of democracies, like in the US, in France, in the UK, they were actually built to select competent elites to act on behalf of the people and not asking people what they think about this law or this decision all the time. That was also related to the fact that information was scarce, difficult to send, etc. But how do you see that model today, this ideal of a representative democracy, now that technology and networks have changed the potential scale of participation? Is it still relevant? Or should it be a direct democracy with people participating all the time? You see my question, right?
Audrey Tang:
Yes. I think in the age of whores and pamphlets, of course, there is no way to broadly listen in real time. So the designer of modern representative democracy did what they could using
Audrey Tang:
the materials they have. So you can, in a sense, scale trust by delegating one way up to a representative. And it's not wrong. It's just incomplete now. Because today, technology allows us to practice this trust that's not just delegated up, but rather distributed horizontally.
Audrey Tang:
Because the idea here is that the peer-to-peer modes, the peer-to-peer learning, is the default way when people get exposed to the communities online. Mind, people by far likes to learn about the feelings, the thoughts, and so on, of people who broadly are like them. So like in the same community, instead of people who are elites, who are unlike them. And actually, the more disconnected you are, the less likely that you get the attention from people nowadays. Most people now want to look at a comment section, at a group chat or things like that, instead of just listening to the one minister or one professor or one chief editor dictate what the society must talk about. So I don't think the representative model is entirely obsolete. I think it still has a role for stability, but now it must be augmented by not just a vertical mode of trust, but rather the horizontal mode of broad listening and participation.
Julien:
I think this is a very central topic here, because even before talking about how we improve democracy, how we build the tools, there is this belief still that, you know, the people cannot take decisions for itself and that we cannot trust the opinion of the majority. And this is why we need to have specialists, experts, people who know how the constitution works who know, who've been there, who've traveled to the best of us kind of, theoretically, to take the decisions for the for the people, What do you say against that argument, if any? Especially when people seem to don't have a lot of time to think about bigger topics, seem to prefer simplistic consumption of online videos. Yeah, I want to hear you on that.
Audrey Tang:
Well, first of all, there's a lot of effort being put into making online videos. We are making one. So just as a matter of record, consuming videos, remixing them, even in online systems like Roblox or Minecraft, to have a kind of different way to organize community and then consume videos in an interactive format. I don't think there's anything wrong or overly simplistic of that. But I think these are all systems where immense collective intelligence emerge and a lot of creativity as well. So I think that the question here is not why people are spending time on video games and watching short videos instead of, you know, going to talk about big ideas. But rather, how can we make sure that those big ideas are manageable in small chunks so that 10 people in a Minecraft room can have meaningful conversations about it and come up with the common ground or the surprising common ground, the uncommon ground between them?
Audrey Tang:
And so that is the first point. The second point is that once we make sure that people are around other people, when we talk about these topics that concerns them, one example is that in Taiwan last year, last March, we sent 200,000 text messages to random numbers around Taiwan asking there's been a new wave of deep fake fraud on social media advertisements. What should we do about them because we're the most free in all of Asia. Nobody wants the government to censor speech or expression. So they gave us ideas. And then in rooms of 10 video rooms facilitated by an AI system, people just became much more creative, much more generative. They came up with great ideas like requiring digital signature and otherwise display it as probably a scam. Another room came up with the idea of full liability, make Facebook, if they post an ad, the full damage if somebody gets scammed $7 million.
Audrey Tang:
Another room said, if TikTok doesn't play by our rules, we slow down connection to their video, but don't censor them, and so on. So in groups, people are very creative and they can entertain this kind of complex systems, big ideas, and so on, supported by the diverse expertise of lived experience of other people. It is only when pulling them individually that you find this inability to grasp with complexity. So maybe don't do that.
Julien:
Well, I guess fundamentally, the battle that is still on when it comes to thinking through the possibility of rebuilding democracy is also exactly that. Some people believe that, collective intelligence, if well organized, can be efficient enough, you know, to make it possible to let the people decide for themselves and basically, you know, govern a country if well organized. And we will go through the question of how can it be organized. And some other people think, still think the old way, which is, no, we need to have people deciding because people are stupid. And I guess the only way to prove them wrong is to say, okay, this is how it works. Exactly.
Audrey Tang:
Yes. And also, I don't think either side is necessarily wrong. I mean, last March, when we had that deliberative conversation, people did vote, and it's 447 people, statistically representative of the Taiwanese population. So their core package where more 85% of people supported and the remaining 15% kind of can live with it also, got still passed to the parliament. So we did not skip the MP. The MPs did deliberate on the core packages and the law was passed in May and July. There's two laws respectively. And then this year in Taiwanese social media, there's just no deep fake ads anymore. So I think the MPs need to see that this way of polling, polling by groups, generative polling, is actually very effective in defining what people want, as well as discovering people's experiences. But the actual work in coding the laws, in developing the laws and delivering the policy, that is still the job of the MPs. We're not skipping past that. So both sides can have something that goes with it.
Julien:
Super interesting, especially coming from France, where it's super top-down, you know, system. It's not like Switzerland when people get to vote often or the places where you ask people opinions and solicitate them to understand what's going on on the field. So I want to understand how it works in Taiwan. And also, I want to go back to your experience, how it started, because you have actually lived through an experiment, you know, that tried to answer all these questions in real time, like, how do we do this? What do we build? So, I want you to tell us the story. I know a little bit of it, but to be frank, not the details, and we don't have the time to go through all the details, but I suppose that most of the people that will be listening to this don't know anything about this. So how did this begin for you personally? And what was the atmosphere back then? How did the idea of digital democracy start to take shape?
Audrey Tang:
For me personally, it was in 2014 with the Sunflower Movement. In March that year, the government was about to pass a trade deal with Beijing in secret. And in response, half a million people occupied the parliament nonviolently and the streets around it. And President Ma Ying-jeou's approval rating was just 9%.
Audrey Tang:
So the system was very broken in a country of 24 million. Anything the president said, 20 million people against it. So I was a civic technologist with G0V or GovZero community. And our motto is to fork the government, which means that if the government doesn't work well, instead of writing it off, we keep what's there, like the information, the code, and so on. And then we steer it to a different direction. And so democracy for us is like a social technology that can upgrade, like semiconductor design. Every half a year or something. And so it was an emergency response. We brought internet cables. I personally brought like 350 meters of cables to the occupied parliament's projectors, open source tools, and so on. And then the half million people on the street and many more online can gather in the groups of 10 or so, as I mentioned, in person.
Audrey Tang:
And so when they do that, the civil society organizations, there were more than 20 of them, talk about specific aspects of the trade deal. For example, people talk about Huawei and ZTE and about Trojan horses in our then-new 4G telecommunications system, and so on and so forth. And so turning that anger and anxiety into like a geothermal engine, a coherent set of ideas co-created. And so digital democracy for me was born from this peak experience of mass real-time digital listening. And after three weeks of Occupy, the head of the parliament basically said, well, the crowdsource ideas are better than ours, so let's just go with them.
Julien:
It's difficult to understand, I think, for people coming from Taiwan in that era, how this could work. Because i guess there is a specific context to taiwan where at the same time you have a lot of people that are very literate about technology about what we want to try about hacking about democracy and what we want to build and on the other side i guess so on the government you have people that were willing to accept to try new things what was specific you know when you look at European countries and you know what was specific to Taiwan.
Audrey Tang:
Well, at that time, I think people do want some way to solve this deeply entrenched polarization. And Taiwan, according to VDEM, was the top target for polarization attacks in the past years. So we had the most intense attack that tried to look at each of the divides in our society. It could be urban rural, it could be ethnic, it could be spiritual, it could be gender, whatever ideological differences. One side want to free China, another side want to be free from China, so on and so forth. And so the polarization attacks, paradoxically, made the Taiwanese people much more resilient. And when we came up with ways to depolarize through broad listening, everybody is collectively tired of peak polarization 10 years ago, so that they're really willing to make it a try. And then after a hundred or so times of collaborative meetings using this method, anyone with 5,000 signatures online can force a response and deliberative conversation. By 2020, the approval rating of President Tsai was over 70%. And that was the year of pandemic 2020. And we only lost seven people to the pandemic. But we did not lock down any city throughout the pandemic. Wow.
Julien:
Okay, and the fact that the democracy is quite young, and the internet and all these technologies is quite young, I guess also that made it easier to try to reinvent things, compared to very old democracies where people are used to stability and the way things work.
Audrey Tang:
Well, yes, I think unless there is an acute harm that is felt by the people. And for many, the pandemic was that acute harm that forced a rethink of how the old system, which works, as I mentioned, in a number of years and at most in quarters, how they can or really cannot respond to a virus that literally mutates in the matter of days, right? So the mismatch between the speed of the threat and the speed of the response system forces everything in many countries. Japan, for example, did not have a digital ministry or a digital agency. But during the pandemic, everybody was like, oh, Taiwan can rationing out masks in really record speed, can do contact tracing without any privacy infringement, can do efficient allocation of vaccination, and also humor over rumor so there's not even a anti-vax fight. And the Japanese people are like, okay, let's have a digital agency so that in the future, if the pandemic hits again, we can have that sort of rapid response.
Julien:
So that was what you're saying about the good management of the pandemic, the fact that there was little tension. That was related to what you had put in place before? Can you go through that so that people understand how it was organized and why the conversation was managed, decisions were taken?
Audrey Tang:
Definitely. For example, GovZero, which is about forking the government,
Audrey Tang:
Registered this website, g0v.tw, and the government services, which is something that g-o-v.tw, if you change an o to a zero, you get into the shadow government made by civic technologies that is always open source and works sometimes better. And because it's all open source, if it does work better, then the government would merge it in. So very early on, GovZero did visualization of budgets. And then the visualization of air quality. And we, in the GovZero, work with teachers in all schools, primary and secondary, to put what's called air boxes, which measures PM 2.5 and other air qualities and write them into a sheer distributed ledger so that people can sense make about air pollution together. Later on, the same model has been applied to water pollution, to noise levels, and so on. And so early on in the pandemic, there was a shortage of masks. And almost immediately, this shared trust built by the civil IoT project about the map of air pollution and other water qualities get repurposed immediately to show where around you still have masks available in the pharmacies. And there's more than 100 Musk map visualizations
Audrey Tang:
Various sectors in our society. And it immediately triggered a better policy because, for example, one of the opposition MP was a data scientist, and she asked the minister that open street map community showed that while you optimize that each mask is widely available to the person near them with the same distance, it is actually not fair because in urban area, the same 10 kilometer is like 10 minutes, but in the rural area, you have to wait for the next bus. And therefore you have to wait an hour, if not more. And so it's actually deeply unequal, even though on the map, it looks like it's equal.
Audrey Tang:
And then the minister at the time said, oh, you know, MP Gao, you are a data scientist. You have exactly the same real-time, transparent, open information as we do, because we update the data every 30 seconds. So teach us what to do. And so she then suggested a better way about pre-ordering and so on.
Audrey Tang:
And then we implement that Thursday. So not even a week after her interpolation. So the point here is that the shared real-time data turned what could be a zero-sum game of political opposition into a positive-sum game of co-creation, again turning the geothermal heat into energy. And that also meant that in Taiwan, from very early on, the virus did not have a basic reproduction number above one because people are masked up. And we also turned the heat of polarization. Some people say only N95, the best masks, are useful. Everything else is placebo. And some people say it's airflow, ventilation, a mask hurts you, N95 hurts you the most. And then we also depolarized that
Audrey Tang:
By saying, oh, look at this cute Shiba Inu. She's putting her paw to her mouth, saying, wear a mask to protect you from your dirty, unwashed hand. And so if you like to wear a mask, I don't like to wear a mask. You're just reminding me to wash my hand. What's to be angry about it, right? So with the repeated application of broad listening and creating uncommon ground, the surprising common ground, we de-escalated the anxiety around the pandemic and through humor over rumor made the resilience possible.
Julien:
So just to understand, the Taiwanese society was polarized at some point. The politics were polarized like everywhere else. But you came in with a lot of people and we can go through that. You managed to work on the polarization through driving people to a platform, sharing information so i guess one of the key challenges was to drive people to a common place now to to to make people use your platform and to make politicians recognize that okay this works just to understand you know how you come from something polarized to something that bring people together concretely, you know, what are the tools, what are the platforms, the challenges?
Audrey Tang:
Sure. As I mentioned, there are more than a hundred mask map visualizations. And so it's not a single platform. It is rather a protocol, a way for platforms to talk to each other, the way for pharmacies to report their real-time stock inventory of masks available. Another protocol for venues to post out random numbers and a QR code that if you scan, it sends that 15-digit random number to 1922, to your local telecom. But then the venue just sees that you have scanned the code and sent the SMS. It doesn't learn anything about you. It doesn't even learn about your phone number. Your telecom operator knows nothing about the venue because it's a random number, and the state learns nothing. But then it is still able, then through the cooperation of all these telecoms in Taiwan, do exposure notification that has served us very well to Omicron. So it is not a single platform that people have to be on. It is a protocol for a platform to talk among themselves.
Julien:
Okay. And you were part of the government when you were doing this?
Audrey Tang:
Yes, I was the cabinet minister in charge of this.
Julien:
So, okay. Well, when you think back to the early moments, you know, people coding in cafes and occupying parliaments and collaborating online, what stands out to you about that period? What did it reveal about how people organize and make decisions together when they are trusted? You know, the key learnings on that. Maybe something surprising that you hadn't anticipated or some people hadn't anticipated.
Audrey Tang:
Yeah, I remember the Taoist saying to give no trust is to get no trust. So instead of asking people to trust the organizers or a minister or the government, the organizers need to radically trust the people.
Audrey Tang:
And that is done through broad listening.
Audrey Tang:
So it is not actually a top-down broadcasting. team set up. We had literally thousands of civic journalists who have these live stream cameras and was reporting from every corner of occupied parliament. This not only made violence easy to spot and therefore almost unheard of in the occupied parliamentary area, but it was also a I trust from the organizers that each self-organized team of people of maybe just 10 or 20 people can have contribution to the trade service deal. So people in kind of ad hoc way had live stream conversations on the record about some aspect that they care about in the trade service deal. And so the protesters are not just protesting, they were demonstrating,
Audrey Tang:
They were putting up a demo, a showcase of how to deliberate better than the MPs who did not really deliberate about a trade service deal. So this self-organization applies not just to policy discussion, but also sanitation, childcare, translation services for international media, and so on and so forth. So it showed that when the organizer trusts ordinary citizens with full information and full access to the tools to organize, they don't descend into chaos. They rise with creativity to meet the moment.
Julien:
Well this is so remote from the way we operate here it's uh it's impressive and i guess and correct me if i'm wrong i would like to know more about this story but, you were a minister so you were doing your stuff you know like building the platforms the digital part but i guess you were having conversations with other ministers and to explain, what you're telling us the fact that we need to trust people that we need to try new things and i suppose that must have been quite important also in the process no yes.
Audrey Tang:
Uh and i had each minister appoint a team of participation offices or peos and together they form a network and Almost all of them are senior public servants, so they're politically neutral, as I am. I am not a party member. I've never attended any party rally.
Audrey Tang:
I'm non-binary in terms of not just gender, but ideology as well. And so the point is that we are a trusted, credibly neutral team of career public service that together determine how to upgrade democracy and listen to people better. And this is crucial because it also requires ministers to think somewhat differently instead of thinking, engaging the public as something that is risky or something that is costly. The participation officer network focused on making each case, designing each case, such that it actually reduced the risk by pre-bunking against polarization and also reduced the cost because they come up with ideas that are generally cost-saving. And for example, one of the early wins is reducing the filing of the personal tax in 2017 from three hours to three minutes.
Audrey Tang:
And everybody loves that because even if you're a minister, you also file your tax, right? So these kinds of wins, that proves that there's not trading off of risk versus time, but it both save time and reduce risk, I believe is really the crux in getting the buy-in from every other ministers.
Julien:
Was it hard to convince people? Because it's really not about just technology here, it's really a different way of approaching the way of governing.
Audrey Tang:
Yeah, I think we are really helped by the fact that President Tsai Ing-wen, when she was inaugurated in 2016, said in her speech, and I quote, Before, democracy was a showdown between two opposing values. But from now onward, democracy is to become a conversation between many diverse values. So basically, depolarization was part of her platform. And having the ministers joining the cabinet on that platform, I think, really helps. And last year, we have seen around the world, I think half of the democratic population went to the votes, right? And not a single ruling party retained all their seats. They all lost some seats. And many of them not to the traditional opposition party, but rather to the more extreme part of the political ideologies. Ideologies, which is why I think now people around the world in democracies are at a time like Taiwan was in 2016, to collectively say, we're tired of polarization and really ready to move on.
Julien:
Well, it doesn't seem to be the case in France right now, to be honest. We're still very much at the polarization level and still one side against another. And I want you to go through that, to go through the process of how democracy can change and learns. Because what you're describing sounds like a living organism, a democracy that breathes, that evolves. Thanks to smart politicians, but also thanks to a smart group of people who are willing to build connections. Connection is something that is important for you. So if democracy is alive, it must have a way to sense, to process, and to learn. So in your opinion overall, what allows society overall to stay intelligent and emotionally coherent to progress? What are the conditions and the risks there that we can observe?
Audrey Tang:
That's a great question. First of all, I think you need to shift from control and command into a shared space. I'm literally looking at my keyboard as I say this shift, control, command, space. And the important thing about this space is that people need to be speaking toward a shared common topic, rather than just dunking each other. This is crucial, because on some spaces, what I call anti-social media,
Audrey Tang:
People dunk, retweeting a person and adding some words to the person, and making everything that could have been a bridge, kind of burn that bridge, and make it about personal attacks, about winning the argument, right? But on the system that we built, like Polis or the join platform, there's no such affordance. There's not a retweet button, period.
Audrey Tang:
There's not even a reply button. There is only upvote, downvote, post your own idea, see the visualization of where you are among the clusters of people, and see the bridge. So this breaks the trolling dynamic because there is literally no way for trolls to grow. And so it shifts from trying to dominate one another, control the narrative, command the narrative into simply a shared space. And the space is larger. I mean, the space key is literally larger. So the shared space is the first thing. And the other thing that is important is to also see that there must be a pre-commitment to the surprising consensus that emerge from the space. I call it the air cover, right? So as a minister, for example, I can say, I don't know what the 447 people will come to terms to fighting the deep fake scam. But I agree that whatever they agree on, I also agree. And I will do my best to convince the legislators of the wisdom of the crowd. If I don't offer this pre-commitment, or even worse, if I offer and then I default on it, then no amount of space making would work. So you need a good space for social conversation and then you need the air cover.
Julien:
But I guess that's one of the key challenges because, for example, that experiment was tried in France a few years ago when it comes to ecological matters. You know, like a group of people were selected randomly and they worked for weeks on making proposals that the president promised to accept in the end. You know, whatever you got, you know, I just have three jokers. I can say no to three things but i will accept the rest and then nothing happened so of course the damaged the trust so my my question my question related to that is, all the system that you're talking about that you build the social system that makes it possible to avoid trolling still exists next to facebook next to tiktok next to other platforms that are built, and trolling that are built on the fact that you will have people fighting against one another. So how do you scale that? How do you attract people and convince them to stop going on the other platform to discuss that and to use your platform instead and to participate in a different way?
Audrey Tang:
Yes. What I describe is not a platform, but rather a protocol, right? A protocol for identifying the divides and also then identify the bridges. And that protocol do not need to be run by the government only it can be run by the social media companies so it's not just the national conversation on climate or national conversation on end of life and so on but rather just everyday conversations on Twitter
Audrey Tang:
Twitter when it was still called Twitter had this idea called community notes previously called birdwatch and what it does is is that for each viral post, you can flag as it is needing more context. And then anybody can contribute with context, saying that, oh, that's actually wrong, or that's not like that, or it's missing important fact, and so on. And then those posts with the notes attached, the poster cannot remove the notes that's attached to it. So it serves as a kind of balancing context. And then the notes are selected, not just by counting the votes, but rather by counting the agreement between the people who would otherwise not agree. So for each note, you see a left wing who support it, and the right wing, which doesn't support it, support their own notes, which the left wing also doesn't support. It's very polarized, but there's a few percent of the notes that both sides consider helpful, the upwind, and these nodes are then featured prominently on x.com. And now this algorithm is taken as a default now on X, but also in YouTube and on Facebook threads and so on. So our protocol, the POLIS protocol of the bridging system, is now the default.
Audrey Tang:
And in the US, they use it to entirely replace the institutional fact-checking. They now rely on this horizontal fact-checking. And just last week, my x.com account enrolls into this beta test of using the same bridging system, but not just for the notes, but for the posts. So the main algorithm to determine the recommendation for the lines that I see every morning if I open the X.com app is now powered by Grok, who figure out the bridging potential of each post and show it to me. And that really changed the atmosphere. I feel a very low polarization per minute once I enroll in that alpha test. And Elon said that before end of the year, you will be able to talk to Grok with your overt preferences so that it will reorder the recommendation engine based on your written or spoken preferences. And so that is a concrete way to make the antisocial social media more pro-social. And we're also building the Green Earth project with the blue sky ecosystem to do the same, but in an open source way.
Julien:
So what you're explaining to me is that currently Musk and X are trying to test new models to limit polarization.
Audrey Tang:
To depolarize, to de-escalate, yes.
Julien:
Okay, okay. Well, that's very interesting because it's quite counterintuitive for most people, I guess, that these platforms are trying to.
Audrey Tang:
Sorry? Yeah, it could be the redemption arc.
Julien:
Yeah, maybe. I guess from what you know, all these platforms are aware of the issue, looking into the issue, despite the fact that it could potentially lower their revenues or lower attentions?
Audrey Tang:
Well, first of all, I talked to many of these people, including the person that invents the like button. And I don't think they meant harm in the first place. They probably just wanted more connection between people, and they naively thought more connection automatically results in better relational health. It turns out the internet is not just good for swift trust, but also swift distrust.
Audrey Tang:
And that's a surprise, a genuine surprise to many of them. So to your question about revenue, actually, if people want to maximize division, they don't need other people now. They can just go into the hamster wheel of AI slop. AI can generate divisive content at industrial scale without involving actual people in those videos. It's happening. Right, it's happening. And so I think what people are seeing now for the first time, for many... Is that this peak polarization, peak addiction is exactly like junk food that hijack our reward system. And the hamster really has no steering power over where the wheel is going. And therefore, relational health is back.
Audrey Tang:
And people really want systems that make relational health the primary thing it cares about and is willing to pay for it, which is why we're seeing much more renaissance of the subscription-based models,
Audrey Tang:
Because people are willing to pay month after month toward an AI system that coheres and depolarizes their social relationships. And that also means advertisers are also catching on the fact that if you can depolarize and have surprising common ground and associate your brand with that common ground, then it's actually more newsworthy and much more of a glimmer, not a trigger, which is negative. A glimmer is positive associated to your brand. And then you replicate something more like watching Super Bowl advertisements together, which is people who would otherwise disagree now share a moment. So again, the advertisers are now figuring it's actually better business if they sponsor this kind of content. So I wrote about this at length with my co-author at a paper Pro Social Media, but I really think this year is the year that Pro Social Media takes off on the revenue scale and also is reducing risk because people are now seeing the antisocial harms of the hamster wheel.
Julien:
Yeah, that would be interesting to go back to that in DNA. I won't have several questions on that, but I want to go back a little bit to your experience with Taiwan at the government level. Because one thing is that many people see emotions basically as noise in politics. And I want to understand what happens when we treat emotions as part of the democratic signal instead. As information that we can learn from? That's something you worked on, right?
Audrey Tang:
Yes. And also, I would say that the emotions, especially around outrage, many people fear that because many people feel it will only lead to populism. But if you have a good geothermal engine, you can be more popular than populism. And then the populist will have to reform to be more moderate. And so staying away from outrage and chaos because that was just a symptom of not being heard.
Audrey Tang:
Polarization is what happens in a system where the politician broadcast and the only way to get their attention is to be extremely loud so that they get overshadowed by your voice despite them holding the megaphone, right? But if you have a geothermal engine that can get the chaos in and have coherence emerge, then people do not need to shout at each other. I personally practice this. I call it troll hugging. So that if anybody in line trolls me, make a thousand words attacking me personally, toxicity or whatever, I use a language model to find the five words in that thousand words rant that actually can be construed as genuine and constructive. And I reply only to the five words, very sincerely. And so the idea here is that people are complaining, ranting, making parental attacks because they feel ignored by the system. And if the system can listen to them, which if I have to do it manually, it's a lot of emotional labor, but hopefully I have released enough public domain material so that all language model can do that for me exoskeletally. So it helps me to hug the trolls, even though the trolls are kind of heavy and then I can lift their weight.
Julien:
Again, a tool and protocol, I guess, super interesting how you can think of technology that can help us through that, getting through the noise. And I guess also one of the ideas of plurality, plurality, not the word, what you're building. What it suggests is that democracy depends on our capacity to hold multiple truths at once, not just listen to one, to other's opinion, but to hold that nuance. How can we cultivate that ability that seems very hard today to cultivate both individually and collectively through maybe protocols or habits or whatever?
Audrey Tang:
Yes, I think there's a lot of attention now being paid for AI systems to work not as like individual dyadic relationships, like a personal tutor or personal slave or whatever,
Audrey Tang:
But rather as a team coach, rather as a facilitator. So it would help, for example, one side thinking about climate justice from a rights-based lens, generational rights. And another side may think in a biblical sense, creation care, and so on. So neither side, of course, want to adopt the framework of the other side. But if you have a good enough facilitator, then they can interpret doing social translation, letting both sides know that they're talking about the same object-level policies, just with different frames, right? So this kind of translation is very exciting because then it shows people that AI do not need to be the kind of addictive intelligence that keeps you glued to the touchscreen. It could be a kind of assistive intelligence so that even though it is a lot of emotional labor for me to hug each troll, once with the exoskeleton of language model, I can hug them with ease. And for the two communities to have this bridge-making, facilitating language model, they don't need a trillion parameter model, use a lot of energy,
Audrey Tang:
Memorize the whole studio Ghibli repertoire, doesn't need to do that. A very small model, the safeguard model, recently open sourced by OpenAI and our organization, Roost, can run on your laptop, and it can do that translation very well.
Julien:
And that's released already?
Audrey Tang:
Yes. So it's called GPT-OSS Safeguard. And so what it does is it's an open version of the GPT model. It can run on your laptop. It's running on my laptop right now. And it takes a policy document, not as law, but as kind of code of conduct. So you can say, oh, this is our community. We prioritize creation care. You need to come up with biblical citations, so on and so forth. And then it upholds the communication standard around that kind of discourse. And another community can have a different policy and a safeguard because it's like personal computing. It runs on the community hardware. It just safeguards the conversation around that particular community instead of like chat GPT that listen from all over the world and try to make some sort of universal rule, which is very difficult, probably impossible. And then those two safeguards can talk among themselves using another open source model, another safeguard model that facilitated the translation between the two.
Julien:
So you could imagine plugging that into any type of platform, even being run by governments, to facilitate ideas from people and having this participatory dream that you talked about. I guess this is the end goal, no?
Audrey Tang:
Well, that is a very good outcome. And in fact, the infrastructure to run the safeguard model called Osprey is donated by the Discord platform. And as you know, in Nepal, they use the Discord platform to do what you just said, which is to choose an interim prime minister.
Julien:
I didn't know that. Can you tell us a little bit about this initiative?
Audrey Tang:
And it turns out in Nepal, the people who started social movements online uses the Discord channels to connect. And when the time comes to choose an interim prime minister, they just cohered online using this kind of digital democracy platforms. And Osprey, now donated by Discord, is now being taken up in production now by Blue Sky. And within Roost, there are also, I think, Roblox and Notion and so on. So this kind of mid-sized platforms are, I think, very good polities, even though they're not like sovereign Westphalian countries, they have enough of a governing complexity so that if they want to share decision-making power, currently mostly around content moderation, but you can easily imagine expanding it to other parts of decisions.
Julien:
Super interesting. Well, I mean, Taiwan's democracy, as you said, has evolved under constant pressure from disinformation, cyber attacks, and attempts basically to influence public opinion, you know, in part from across the strait, I guess. The same is happening all across Europe, and especially in France. You know, we have a lot of Russian trolls, and uh and and i'm saying this to all the the trolls coming to see the video please comment you know you're helping me growing my algorithm it's fine yes.
Audrey Tang:
We can hug
Julien:
You we love you we hug you yes thanks for being here uh but i want to understand you know how did you navigate that tension between openness and security and security and and what did these uh experiences teach you about defending democracy without becoming defensive.
Audrey Tang:
Yes. I think the lesson we have learned is that staying secret about policymaking creates a void where the polarization attack can grow. And the only way to fight that is not by debunking it, which doesn't work and only feed the conspiracy theories, but rather by pre-bunking it through radical transparency. So when everybody, including the opposition party, can see every 30 seconds where exactly are the mosques, there really is no room for conspiracy theories to enter.
Audrey Tang:
When people can see transparently that the preferences of each age bracket in each district about each brand of vaccine, there's really no point for the anti-vax fight. People instead compete about which brand is better. I got all four brands just to make a point. But the point here is that radical transparency writes the air instead of just people reading the air. It's a Japanese expression where people individually read the air trying to guess what other people think. But actually, with radical transparency, you can write the air to make common knowledge so that people all know what's going on. And then in that environment, people actually feel more secure because we have cognitive security that not only I know this is happening, I also know that you know this is happening as opposed to a culture of secrecy, which leaves people second guessing each other and they all suffer from cognitive insecurity. Like, is this actually happening? And this actually is very toxic to democracy If we cannot even agree what's actually happening, how can we do sense-making and make decisions together?
Julien:
I guess the tricky starting point is to be able to trust the government website that is offered to you that's well it's a.
Audrey Tang:
Ledger right uh the environmental sensing project airbox started by people distrusting the environmental agency's numbers because the measurement stations are too few and far uh in between and so people feel that air pollution is really bad but then the environmental agency was saying, oh, actually, it's not that bad, right? So which is why people measure it themselves with the primary school teachers and so on. And then because of the distributed ledger, the government cannot fake the numbers and override the numbers. And people cannot override the numbers of the government either, and the industry also. So none of the stakeholders in this distributed ledger have the way to go back in time and hack the numbers because it's secured by distributed ledger technology. And that, I think, is one of the very good uses of blockchains as a notary. So it's not a smart contract or anything. It's just a public notary that nobody can tamper with.
Julien:
Yes, and I guess... People need to understand how that works and what is a distributed ledger. So you need to have a basic education on all these things. Is it something actually that you had to work on to educate people to technology, to how the system works? Because I guess that's one of the key challenges, right?
Audrey Tang:
Exactly. Yeah, in 2019, we changed our national curriculum on what's called a literacy-based curriculum, like data literacy, media literacy, which teach people how to consume information, critical thinking, and so on, individually,
Audrey Tang:
To what's called a competency-based curriculum, which is about people co-producing information, sensing the air together, fact-checking the presidential candidates as they're having a debate, so on and so forth, starting petitions to go to school one hour later, because one more hour sleep gets you a better grade, as proven by science. They did get that change, and so on. And so getting people into the habit of producing information together, not only depolarize those young minds, but also it is probably the only way that people have the motivation to learn about distribute ledgers in the first place, because then they really make an impact on whether their parents and grandparents go to hiking that day, or do they stay indoor because of air pollution, right? So it really made the entire nation listen better to young people's contributions. And the young people became like reverse mentors to the older generations. And by showing that digital competency can work for the societal good, not just the individual accomplishment. So by 2022, our middle schoolers are now top of the world, according to the International Citizenship and Civic Study, when it comes to civic knowledge, the civic muscle, that they feel they can already change the society even before they turn 18.
Julien:
That's astonishing to me. So middle school, people like 12, 13, 14 years old learning that. Everyone, everyone looking into these questions. And that was your minister that did that?
Audrey Tang:
Well, I was part of the curriculum committee before joining the cabinet. So that was my pre-cabinet job. And that was already after AlphaGo. So we reasoned that if we keep the old curriculum and train people in robotic routine skills, then when there were seven, they may feel, okay, I have to learn this. But by the time they're 18, which is around now, they will hate us because robots will take their job, which turns out to be true, right? So instead of teaching people the skills that will get automated eventually, we only teach three core competencies, which are curiosity, collaboration, and civic care, and nothing else. So if people really are good at the civic muscle, at those three core competencies, people then see robots as friends because they're not in danger of getting their meaning taken away from automation.
Julien:
You mean you mean nothing else it's it's but i guess they learn many other things but it's it's the core of the program you mean this is shaping the rest of the of the subjects of the topics that they learn just to understand you know nation.
Audrey Tang:
Works so instead of like how well do you memorize the standardized answers we measure instead how much creativity you show in teams in working on a novel problem that benefits your community, right?
Audrey Tang:
So the way examination works also changed in Taiwan. I think the point here I'm making is that learning other things is an instrument in fostering curiosity, collaboration, and civic care. It's not the other way around. It's not that you pay fake attention to your neighbors just to maximize your score. The maximizing score part really almost doesn't matter now.
Julien:
Okay, so you reshaped the school. You reshaped the way the government works. You reshaped protocols and tools. I mean, now that you have stepped back from government, how do you see what remained after that experiment? You know, what are the coalescence? Do you think the rest of the world could learn from it? And maybe what is frustrating to you that is not sticking?
Audrey Tang:
So first of all, I think I was always at a Lagrange point between governments and social movements. A Lagrange point is a point between, say, the Earth and the Moon, where it's not falling into the orbit of Ida, right? So it gets gravity from both sides and acts as an effortless channel, wu wei, in a very Taoist way, that can translate between the two perspectives. and I'm still there. I think as a cyber ambassador at large, cyber means steering. So this is not about accelerating tech, but it's not about stopping tech either. It's about steering tech so that it becomes tech in the loop of communities instead of human in the loop of AI, as we talk about. So I think the steerability was the main lesson that I want to share with the world That is to say, tech is not something that happens to us.
Audrey Tang:
The Silicon Valley way of building a superintelligence that sees everything, does everything, is a lack of imagination. We can actually build tech that serves the community's needs. Not all tech need to be super progressive. Many tech can be communitarian. There could be civic community AI systems and so on. So that's the main lesson. And I think people are very eager for that lesson now because many people are feeling a sense of disempowerment when their children, when their friends, their family members kind of fall into this reality distorting relationship with chatbots and AI companions and so on. So a way for a community to take back the steering power, I think is very important. Now, what was sticking in Taiwan, again, was this geothermal engine. So more and more conflicts, even the very old conflict of the free China and free from China divide in Taiwan, arguably the most core. Earlier this year, we've seen a rally where the two flags actually flew together, which never happened before, against authoritarian polarization. And so I think I'm really proud of that. It shows that it's not me personally at the helm that makes the difference. It's the protocol, the geothermal engine, that makes the difference.
Julien:
I mean let's um let's dwell a little bit on that and what's happening outside you know what is the new what are the new front lines kind of you know because we now see competing models you know control in china you know a little bit of chaos in the us and a few civic experiments you know in between i mean the chinese model shows how technology can also become an instrument of social control you know in some in some ways you've got you know the american model it's often turns into a kind of infrastructure of uh more manipulation and profit even though you know we can go back to that to see you know how it's progressing but for now this is mostly what what's happening what do these extremes reveal about the deeper relationship between technology and democracy overall you know i want to to see to understand how you look at at the big picture of that tension.
Audrey Tang:
Well, I mean, whether it is a hamster wheel that serves the command key or whether it's a hamster wheel that serves the control key, I think these are fundamentally about the lack of option, the lack of escape, which are also keys I'm looking at on my keyboard. So the command and the control impulses that people strongly associate tech to is, I think, really a philosophical lack of imagination. People can only imagine two ways for technology to operate. One is called a utilitarian way, which is to maximize some number. And the ontic way, which is to follow some command, follow some rule. And nothing else, right? And people are like, oh, we should put the brakes, stop AI, follow some rules. Oh no, we should accelerate, maximize this number. But if there's a car with only the gas and the brake, the car is not going anywhere. Well, maybe falling off a cliff, which isn't going anywhere, by the way. The point here is the steering wheel. And instead of viewing citizens as a resource to be extracted, it, as you say, about obedience, attention, data, or whatever,
Audrey Tang:
Then the struggle here is between those addictive models and assistive model, which builds the space key and treats citizens as agents to be empowered,
Audrey Tang:
Gradual empowerment instead of disempowerment.
Julien:
How do you see things happening in Europe? I'm not sure how much you know about the regulations that are being put in place, the approach. I want to have your point of view on the European approach to these questions.
Audrey Tang:
Well, I think the European approach of shaping the market through radical interoperability, I think it's really, really good. The Digital Markets Act, for example, say that if you're an instant message company and your client wants to switch to a competitor, you cannot deny them access to the old address book. If I switch from one telecom to another. I should be able to take my number with me so that people call me. I can still pick it up without having to switch to a different number. And so that is really powerful because then it shows building information superhighway must always come with an off-ramp so that people can choose between different offerings and forcing them to compete on merit instead of on how quickly can you, through the network effect, capture people, and squeeze them. And so I think that's very good, and I wish it could go further.
Audrey Tang:
The Digital Markets Act at the moment is in another round of consultation about expanding it not just for instant messaging, but for social media as well. Actually, the state of Utah in the U.S. already passed the Digital Choices Act. Let's say starting next July, if you're a Utah citizen, and you want to switch from x.com to, say, Blue Sky or True Social, You can keep your community. The new likes, reactions, followers just flow to your new network. So people do not need to feel trapped in any of those social networks. So there is a future, I think, where it's now the default for just people in Utah, but through the Digital Markets Act, maybe for everyone in the EU. And maybe the Europe can do the euro stack for every part of this through radical interoperability, become almost like a decentralized euro stack. So instead of, you know, the national champion Airbus building the European option for each stack, maybe you can just force the Silicon Valley and maybe Beijing players to become like utilities, where they have to operate in the open, inspectable, and most importantly, freely offer portability among the choices.
Julien:
Some people say, some critics say that all these regulations, especially in Europe, are slowing down innovation, are preventing Europe from being a key player, you know, on the AI race or, you know, data in general. What do you think of that, critics?
Audrey Tang:
Well, I mean, if you have off-ramps, of course, people are going to drive on off-ramps if they don't feel that this destination is where they're going. If you don't have off-ramps, I guess, yes, you do accelerate, but then it leads you somewhere you don't want to go, right? And I think that is the main idea. I am an accelerator fellow, a senior accelerator fellow at the Ethics and AI Institute in Oxford. So our program is about accelerating choice. It's not about stopping innovation. It's about innovating toward where the society wants. And that is probably what the innovators truly want anyway. It was just because of the high PPM, because of the slow feedback loop, because of the megaphone broadcasting only model, the innovators really had no idea what a society really wants. There's no way for them to be truly attentive to the society. So if we figure that part out, I think most innovators would prefer if their innovations are taken by the society as something that a society can actually use rather than something that society gets strapped on and just, you know, keep spinning the hamster wheel just because they cannot get off.
Julien:
Yeah, I'll go back to that a little bit later, but I want to talk about AI a little bit to understand what are the challenges, because AI can be used to expand participation, but also to manipulate perception. And we can see that you know now i don't even know the the share of content that is now produced directly or indirectly by ai but it's basically invading you know internet and it's causing a lot of trouble it's.
Audrey Tang:
Now half of new content
Julien:
Yeah so so it's basically you know killing the some of the i mean damaging some of the bases on which internet has been built for for years so how do you see this tension between kind of empowerment and you described some ways of using AI in a smart way empowerment and control how this is playing out in now and in the coming years.
Audrey Tang:
Well, first of all, I think people are becoming aware of the slop, right? And the propaganda bots, the deepfakes, emotional manipulation, so on and so forth.
Audrey Tang:
So it's almost like the time when the Montreal Protocol around ozone was being signed. Like people for the first time clearly understood that freon, a specific chemical, is the cause of the depletion of ozone. And then people collectively say, let's commit, even though that we don't have off-the-shelf replacements for Freon, let's commit that a few years from now, everybody will use the new thing that doesn't destroy the ozone. So I think we're at a very similar moment. People are seeing addictive AI as that kind of ozone-depleting thing that strip mine, the social fabric, and kill or at least harm relational health. And people are also seeing that there are social translators, consensus finder, personal kami, relational kami, the local stewards, and so on, the assistive AI as a powerful tool for empowerment. So I think the sooner we can set a standard so that anyone who do not use this kind of assistive AI and insist on building addictive AI will be seen as destroyer of ozone, then the better can we coordinate around this. And interoperability and transparency and the safeguard model that lets community bring their own policy to keep AI honest, essentially,
Audrey Tang:
And also to process evaluations like weevil.org and so on. All these are a part of constructing a human loop to bring AI into instead of forcing individuals into the loop of AI.
Julien:
But you know these guys, you know that the people working on taking decisions on some of these topics. And from my standpoint, I see the tension here between, again, as we said before, the profit and the possibility to use AI to capture attention, to make much more profit, to capture a market and doing something else and limit basically, self-limit one's power for the sake of democracy. So I want to talk about this tension specifically and also the conversations that are happening right now and how this is happening. What is the, where is it happening? And is it happening? Is it being discussed seriously? The dangers of AI when it comes to making sense of the real world, making sense of information and for democracies, what are the conversations, what are the protocols that are being put in place to work on that?
Audrey Tang:
Well, it's very much happening, right? And first of all, I think social media, the recommendation engine being the exhibit A of how a parasitic addictive AI can harm the society is now well understood. So much so that like in Australia, they even say that starting end of this year, you have to prove you're over 16 in order to use social media. So that is basically...
Julien:
Good thing, yeah.
Audrey Tang:
That is basically saying, you know, it's like learning to drive. If you don't have a driver license, it'll probably kill people with that machine. And I think it's a good thing, as you say, that a society is now realizing about the harm and also saying this is not about individual choices. If a child or the classmates use some sort of social media, there's really no way for the child to resist the peer pressure. So it's not something that you can leave to individual parents, individual childs to resolve. This is something that the teenagers and their parents need to come together and build consensus and then to write the air instead of everybody reading the air. And so piggybacking on that, we're now also seeing conversations about AI companions or chat bots, providing emotional support. As part of the Collective Intelligence Project, CIP, we run global dialogues every two months, having a worldwide conversation, putting people in rooms to react on each other's ideas around the use of AI.
Audrey Tang:
Consistently, people trust the chatbots 30% more than the company that makes the chatbots. Which is very interesting. It also talks about the importance of the community to then to own those chatbots so it doesn't need a company puppeteering from above. And of those puppeteering chatbots, one in seven people now say they or a close friend of them have reported a reality-distorting episode where the chatbot is leading that person kind of off the baseline reality. And 1 in 10 persons say that sometime in the conversations, they don't have an agenda setting power anymore. It is this conversation is being steered by the AI companion, which is, again, very alarming, right? So I think it's great that policymakers around the world are waking up to the fact and we don't need to, again, wait for 10 years, like in the case of social media, to have such conversation. These conversations, including so-called AI red lines, are being had in many different polities, in state level, in the UN, in many multilaterals.
Julien:
And it's happening also among the companies themselves?
Audrey Tang:
Oh, definitely. Yes. I think the idea of gradual disempowerment, of people yielding their judgment to this all-seeing, all-acting Skynet thing, I think is the number one failure mode because at the end of the day, those innovators do not want to be captured by the robots either. They're also on the human team.
Audrey Tang:
And I think the point here is how quickly can we say, oh, actually, you can build for relational health, that you can still make a refrigerator without destroying the ozone. And it's actually not a hit to your bottom line because people are going to refuse the use of the ozone-depleting refrigerators very soon. So the market incentive need to shift such that people demand this kind of right of portability, of audits, and so on, to such companions. And if they don't deliver on it, then there needs to be a real alternative of community-scale AI so that people can run it themselves. One of the best-kept secrets in the AI industry is that all those large models fit on a USB stick because they have to fit in a GPU memory. So actually, most of our personal computers and some high-end phones can run it as well as the large data center. You don't actually need a data center. So the personal computing movement, the maker movement around AI models, is at this moment still kind of hobbyist. But so was the personal computing in the beginning. It was also a hobbyist thing.
Julien:
So what you're saying is that maybe in the future, AI will move into our hardware systems locally, and we don't depend on big corporations, and we will not be fearing to be listened all over the place. People are working on that, right?
Audrey Tang:
Right. I mean, for the past two years, if you write me an email, I draft my reply using a local AI. It works in airplane mode, right? So I know that my emails drafts are not going out of my computer. Of course, eventually I hit send and it goes out from my computer. But my point being that this whole training process can be done entirely locally.
Julien:
Yeah, okay. We need to make people much more literate about these topics because for now, AI companies are dominating the game. And, I mean, from where I stand, you know, when I listened to Sam Altman being interviewed at the last TED talk, I guess, you know what I'm talking about, it was not really reassuring. It was a bit, you know, vague and, okay, we're trying to build it. We know there are risks, but we're not really sure, you know, how we're managing it. And I know you've been collaborating with OpenAI. Well, you might correct me if I'm wrong. Of course. So, and other global players. So what are the conversations you were having about all these topics, about the fact that we need to decentralize things? And also questions, how do you navigate that space without being captured by corporate or ideological agenda? I guess this is a question that you often have, right?
Audrey Tang:
Well, I mean, I'm not dependent on really any resource from them. And also, I think the idea that OpenAI needs to be open is in their name, right? So the fact that they opened the safeguard model, the GPT-OSS, and so on, most people join OpenAI dreaming to do something like this. It's in their charter. It's in their name. It just took quite a while for them to get to this point, right? So in every large organization, there's plenty of people who are in pockets of good that wants to do good. It was just the market incentives that for a while make it more profitable, I guess, for them to not do good, not to open their models in a safe way, to share the way the safety pipeline works and so on. So my role is to be a translator. Because I'm a technologist, right? I programmed with the Siri team, actually, for six years, bringing cloud service localization.
Audrey Tang:
When I joined the cloud service team, Siri only spoke English and a little bit of Spanish, I believe. And my last project with them was doing the Shanghai language, the Wu language in 2016. So I know how these systems work. And so when I have policy ideas, when I have philosophy ideas like care ethics instead of just utility and deontic, I show them in their own language how to embed those policy and philosophy into their models. So I'm not captured by their agenda. I'm kind of sharing wisdom in their own language, the plurality agenda, how to make broad listening work better.
Julien:
Yeah, and I guess you find a lot of people that are willing to listen to you. It's not evil. If we think of AI as a new nervous system, potentially for democracies, what principles should guide its design so that it fosters empathy and understanding instead of potentially more polarization and fear? Do you have ideas on that?
Audrey Tang:
Yes, definitely. So I think the way to steer such AI systems into the nervous system of our democracy is never from the above, but rather with the people, not for the people. And if you want to work with the people, you can check out sixpack.care, which shows the six pack that such system need to pay attention to. It starts with attentiveness, actually listening to people, not just the popular, not just the powerful, but small underdogs too. And it involves taking responsibility, actually keeping promises. So not vague ideals, soon abandoned, but specific commitments with teeth. And then people should be able to check the process, not just blindly trust the powers that be, but transparency and fast community feedback. And finally, what counts as good, the results must also be crowdsourced.
Audrey Tang:
So it's not metrics that don't measure what we value, but metrics designed with the people. And so then that goes back to attentiveness. So this is the brainchild of the philosopher John Tronto, who coined the care loop, attentiveness, responsibility, competency, responsiveness, and back to attentiveness.
Audrey Tang:
And then later on, she added solidarity, which means as win-win as possible. So not mutually assured destruction or zero-sum games, but rather design the games where all sides are better off. So portability, interoperability, the ability for people to switch between AI systems and also switch between ways to aggregate people's feelings in democracies. That choice is also important because it enforced the competition to the top, not to the bottom of brainstem. And finally, the last pack
Audrey Tang:
Which I call symbiosis, means as local as possible. So not a one-size-fits-all overload, but a variety of local stewards by and for a variety of tasks. And this is, I think, a very potent idea because many people in the West, they see the superintelligence and they think Skynet, like one-size-fits-all, literally. But in Japan, they have this idea of local kami, which is local steward spirits, the safeguarded relational health of a river, of a forest, of a village, in which case it's called a uchikami. And those kamis are in relationship with each other, but there is no overlord that presides over them. So it's like an Illinois-Ostrom idea of subsidiarity. So resolution is on the most local level possible. So these six pack, you're going to think of them as exercising care with the nervous system and have six packs of abs, I guess, and also six packs of beer because it's very portable. But the six pack of care shows then that if you optimize for the relational health and civic care, you do not need to optimize any particular number, but then people still report flourishing in the democratic system.
Julien:
Wow. So, in a way, that's imagining ways of outsourcing intelligence in pockets of intelligence, you know, building pockets of intelligence that can help communities or even individuals, communities, governments, and maybe global issues in a new way.
Audrey Tang:
Yes. It's helping us in the gym. It's like the exoskeleton that help us lift the weights and the trolls better. But it's not replacing human relationships. It's not me talk to my robot, you talk to your robot, our robot go to the gym, they lift the weights. I'm sure it's very impressive. But then our muscle will atrophy this way. So it's exoskeletal, it's assistive, it's not replacement.
Julien:
Yeah, and what could come out of this? Yeah, we have no idea, basically. New ways of organizing, new ways of... Yeah, okay, interesting. Anyway, I'll put all the resources on the website for people to go deeper. You're now building a broader movement around these ideas. Plural technology radical exchange and you even like I read that you were collaborating with Governor Newsom also in California on some topics so you all run the place as you said you travel a lot, how do you see your mission today.
Audrey Tang:
As I mentioned, I'm a cyber ambassador, where cyber means to steer. So my mission is to spread the code, both the technical and the social,
Audrey Tang:
I guess philosophical as well, of how to steer. So translating this geothermal engine for other cultures, including California, as you mentioned, but also with more conservative people like the Napolitan Institute in the U.S., who just did the We the People 250 broad listening experiment with five people from each congressional district in the U.S. And asking them, what do you feel when you hear the word freedom and equality? And using this kind of broad listening tools to help people listen to each other better and discover no matter how ideologically split they think they are. It's all an illusion. People do agree broadly on what freedom and equality means, surprisingly, and also what to do concretely together. So it's also with some conservative think tanks and institutes. And also in the UK as well, I'm working with the National Strategy Project, which is now starting a new institute for the nation. And to get the British youth, especially, but all people, into this exercise of civic gym to talk about, I don't know, climate or housing, immigration, and so on. Again, on a hyper-local level, but then see the surprising agreement across hyper-local levels. Thank you.
Julien:
Do you work with people in France?
Audrey Tang:
Well, many people in France, like make.org, actually have that kind of platforms. And I'm a scientific advisor in their democratic commons, I believe, efforts to make the panorama system first open source, but also training AI models that specifically are auditable and transparent for this kind of democratic exercise of civic muscle.
Julien:
Okay a few a few questions we get into the end i know it's late but um yeah if you if you were a minister of digital affairs in france or president of the eu you know for for a week i mean in a situation of uh implementing change quickly um what would be your first moves well.
Audrey Tang:
I would put in the Digital Choice Act of Utah through the Digital Markets Act framework to all parts of the Eurostack. So this is not just about social media, AI services, companions, and so on. It's about ensuring the freedom of movement, the right to exit from the off-ramp to another on-ramp for all parts of the stack, starting from the compute all the way to enterprise applications, social graph portability, and so on, because that would instantly change the market. It would make all the platforms, whether foreign or domestic, to stop competing on addiction and the hamster wheel and starting competing on care that who can build the best steering wheel and share with the people. So it is a profound power sharing move by sharing power back to the people.
Julien:
Where do you think we should focus our collective energy in the next decade? What deserves protection and what needs to be reinvented? I know you have a broad picture and you connect many dots. I'm very curious to know where you answer that.
Audrey Tang:
Well, I think we need to protect and indeed conserve our relational health. So this is the conservative part of conservative anarchism that I practice. Our time spent in communities is precious. It could be spiritual. It could be sports. Why not? It could be fashion. But the ability for us to sit with our families, with our friends, to walk in nature, to listen to our social fabric, the protection of that relational health is really paramount. Because otherwise, we're not truly healthy. Even an individual that is healthy on a physiological basis is not healthy if it's relationally deprived. And so the time that we spent with each other, either offscreen entirely or in high bandwidth communication as we're doing right now, that I think is paramount. And technology should serve such communities instead of just abstract progress. And so what was the other, what to reinvent?
Julien:
Yeah, what needs to be reinvented?
Audrey Tang:
Well, I think the public spaces need to be reinvented. We need to install a new space key. And instead of the addictive cycle that we've been kind of misusing, the addiction-fueled social media as some kind of public square where it's actually not public at all. It's very private sector driven. we really need to re-imagine our digital participation infrastructure and with good digital privacy infrastructure so we can participate with the confidence that we're not going to be doxxed, that is to say re-identified but we can also participate in confidence that if I say that I'm a resident of the city or I'm over 16 years old I can produce a credential to prove this without revealing any other private information that we have. So this solves the problem of bot or not. And it also makes it impossible for black hat hackers from identity theft and things like that because these spaces would not collect personal data in the first place. All it needs is to verify your claim and then you're in and you're having a real conversation among other people.
Julien:
What scares you the most today? What gives you hope? And the third one, what energizes you the most?
Audrey Tang:
I'm sorry, what was the last?
Julien:
Where do you get the most energy from?
Audrey Tang:
Ah, okay. Yes, what scares me is a loss of imagination. If people stop imagining positive vision, then there is no positive action. If all people can imagine is an accelerating car or hitting the brakes to the bottom, we would never steer away from the polarization per minute and into a broad listening. So I think the imagination, the capacity to imagine and share the imagination is the most important. And it scares me when people say they literally cannot imagine a steerable technology. And what gives me hope, of course, then are the digital natives. The young people have surprised me time and again that they, after living in the high PPM environment, they pre-banked this information for fun, right? They saw through the addictive system and choose to build in-person connections through assistive systems. And in Taiwan, when we made civic muscle, the core educational curriculum,
Audrey Tang:
Many young people then become entrepreneurs in the best sense. They become social entrepreneurs to deliver for the people and the planet and find business model for them. And again, that then reorients the incentive system of capitalism to the best. And so what gives me energy is to listen to the stories of these wonderful social entrepreneurs. I just delivered a closing keynote at a social enterprise world forum in Taipei City and also participate in the weaving of those stories so that one insight gained from, for example, climate accounting and so on can transfer to the social polarization per minute accounting and so on. across many different domains. This ability to cross-translate across many cultures, I think really energizes me because I exist, as I mentioned, as a bridge in the Lagrange point.
Julien:
Interesting. I have two little daughters, like six and nine, and I'm struggling with technology. I don't know what to do with screens. For now, they're very much isolated from it. And I know we'll get into this, but what are the pieces of advice that you have to parents when it comes to teaching technology or dealing with these technologies, social media in general, what works?
Audrey Tang:
Well, what works consistently in Taiwan is to treat those screens as something to be shared. So I think most schools in Taiwan ban small screens. And they use only large screens. But we ensure there's one screen per child that they can share with their classmates and teachers. So it could be a large tablet. It could be a laptop. But the point is that screens are meant for sharing, not addiction. I personally always use a large screen. I turn it grayscale most of the time if it's a touchscreen, so I don't get addicted. Or I use a stylus or voice or keyboard. So to me, I can always put it down. And most of the time, I don't need my phone, actually. I've been walking around without my phone. I don't feel a need to carry it around. And the point here is that once people see a screen as something to be shared, something social, then they expect the applications on screen. To foster the relational health between those two people. Maybe it's a translation service. Maybe it's a game that they can innovate together. Maybe it's your drawing or whatever. But there's no hamster wheel anymore if you are sharing the screen with your child, with your classmates.
Julien:
Super interesting. You're a spiritual person. What's, not forced to answer that one, but I'm curious, what's the meaning of life to you?
Audrey Tang:
Well, for me, the meaning of life is to participate in the co-creation of the canvas. So the generations that are born after me have a wider possibility space compared to the one I was born into.
Julien:
That's the good ancestor?
Audrey Tang:
That's the good enough ancestor. Because a perfect ancestor will foreclose possibilities by choosing for a future generation. A good enough ancestor just published before I perish.
Julien:
Yeah, that's very central to your daily life. I can say that, yeah, in a very concrete way. Last question. If you could offer two books to someone you love, you know, books that maybe shaped your worldview, you know, which would ease people's life or, you know, for them to bring on an island into open creativity, whatever, what would it be?
Audrey Tang:
Well, the first would be the Dao De Jing by Lao Zi, The Wisdom of the Way. For English speakers, I recommend the Ursula K. Le Guin translation. It's pure poetry. You can even find it on GitHub online. It teaches us how to see the flow of the Dao, of the way, and how to act with wu wei, which is effortless action. And if you are interested in this kind of binary, the raw binary of Taoism that goes even more ancient than the Tao Te Ching, then another book is the Yijing, which is not meant to be read cover to cover, but rather is the binary code that defines the Tao, that defines the interactions of the way. And many people use it as kind of a demination tool that literally just randomly pulls from it. There's many ways to navigate it like a maze.
Audrey Tang:
Spinoza, I think, got inspired by I Ching for the binary system. And I think the I Ching, like the Tao Te Ching, exists before people even had this idea of a deontic rule that should top-down regulate the world. It was born in a more ancient time where the relationships exist almost tangible, and the individuals are just vehicles that the relationships kind of inhabit. And reading the I Ching and the Tao Te Ching reminds people, like before we turned four, we were also like that. And so it goes into this prehistoric and pre-language state of the mind, the mind of the flow state. And I really recommend both of them to people. And that's also because I'm a Taoist. These are the foundational texts of Taoism.
Julien:
Thanks so much, Audrey. Thanks so much for your time. Lots of insights. I put the notes on the website as usual, but really, really insightful. Thank you so much.
Audrey Tang:
Thank you so much for the great questions. Live long and prosper.
Soutenez Sismique
Sismique existe grâce à ses donateurs.Aidez-moi à poursuivre cette enquête en toute indépendance.
Merci pour votre générosité ❤️
Suggestion d'autres Ă©pisodes Ă Ă©couter :
État de droit : comment les démocraties déraillent
Populisme, justice, démocratie : décryptage des glissements vers l’autoritarisme
Souveraineté : l’Europe à genoux ?
Défense, énergie, technologie, économie… Les dangers de nos dépendances.
France, Europe, Big Tech : Le prix du renoncement
Thierry Breton raconte les erreurs du passé et les rapports de force d’aujourd’hui.
Hypnocratie : bienvenue dans l'ère de la transe digitale
De la post-vérité à l'hypnose collective
Autonomie cognitive : le dernier rempart (par Tariq Krim)
Protéger l’accès au réel, un défi clé de notre époque