We always edit for clarity and for length, obviously. But I will send the quotes to you so you can check.
Yes, OK. I also have my own transcript. And we don't scoop you. So, after you publish, we publish on archive.tw. I've been doing that for 10 years.
Yeah, you sent me the link to that.
Yes, that's right. Yes.
Is it a way to remember what you said five years ago?
There's that. And also, I think more importantly, because we — I want to empower journalists. My parents are both journalists. And I think in the rush to get scoops, most of the time, the context gets lost. But if the entire context is readily available, nowadays journalists can just use language models to pull the grounding facts. And that's very good for investigative journalism, which I believe is a very important social technology.
No, you're right, because usually when you do research, you always get sort of like what the AI tells you — sources that might not be exactly as good, it's not grounded, right? And if it's directly in an archive, yeah, you have it and it's all there.
Yes, when I went to Sweden last year to receive the Right Livelihood Award — the alternative Nobel — it was the first time it was given to someone from Taiwan, because usually they give it to peacemakers. And so this is the first time, I think, the jury recognised the information space, cyberspace, as a conflict zone. And so it's not just on the beaches where there's a contest, politically, geopolitically — there's also, cyber-politically, this kind of information manipulation. And so peace requires integrity of information. And I really think journalism is —
Yeah, I mean, I'm glad to hear that, because I think journalism as a whole is quite under pressure these days, right? Especially when it comes to how AI is being used in journalism. I think it's all quite new and fresh. And I think it will be very interesting to see what's going to happen in the next one, two years. I mean, things happen so quickly now, right? Who knows what the media landscape is going to look like in five years' time.
Taiwan, as you know, gets free red teaming. Here you have to pay for cybersecurity penetration testing. In Taiwan we get it for free, millions of attempts every day. And so what we have found is that pressure really makes diamonds. Because of this constant testing, we have no other choice but to innovate. Information integrity, the trust shield, if you want, is a big part of it.
Yeah, absolutely. Let's talk about Xi Jinping first and then let's move to technology, if that is fine, of course. Go ahead.
Sure.
First of all, I think the Donald Trump–Xi Jinping summit a couple of weeks ago was obviously big news and there was a lot of debate about that. Many commentators in the West felt that in the language that Donald Trump used, it was quite a sort of new tone of language about China. Some commentators said it feels a bit like betrayal, even.
The what? Sorry?
It even feels a bit like betrayal.
Sorry, which word were you alluding to?
He did a Fox interview and he refused to commit to defend Taiwan. He claims Taiwan is stealing American technologies. I mean that story has been around for a long time. And he literally said Taiwan should cool down. What do you feel when you hear these comments?
Well, I mean, as we were just mentioning, the main thing about Taiwan is not just a chip — some call it a silicon shield — but rather about trust. So, people around the world, regardless of what kind of technology they're developing, trust Taiwanese makers and the cybersecurity-resilient supply chain in order to run the most sensitive computation. So, I think it means that we are indispensable, not just on the supply-chain basis, but rather on how the entire world trusts Taiwan to host the computation. So, I think that is our main posture. It is not saying that Taiwan is a problem, but rather Taiwan is the solution. It is the solution to the trust problem that the entire world, especially democracies, is facing right now.
Do Taiwanese people trust Donald Trump?
I think the idea is to earn trust by making sure that we invest in what's called trusted tech. So, I worked with Keith J. Krach during the first Trump administration. Maybe you have heard of an endeavour at the time called … Yes, so maybe you heard of the Clean Network initiative, because back in 2014 we helped half a million people on the streets, many more online, having a deliberation during the Sunflower Movement about whether we should allow Beijing investments into our cybersecurity, telecommunication, publishing, and so on.
And of course, the people in Taiwan, regardless of their party, felt that we really only want to work with suppliers with functioning journalism, actually — whistleblowing and things like that. So that even if some of the vendors want to put in back doors and so on, their own immune system, the democratic journalism ecosystem, will be able to expose that, and we feel safer doing business with democratic entities. And I think Keith at the time, during Trump's first administration, really took up that idea and convinced many of our democratic allies that instead of just putting all the eggs in one basket with 5G and many other dependencies, it's much better if all the democracies just transparently know which technology suppliers come from a trustworthy supply chain and which have maybe hidden Trojan horses.
So, I think that is now expanding beyond just 5G or telecommunication and becoming what's called the Global Trusted Tech initiative. So, I chair the Global Trusted Tech (xGTT) Standard initiative. And so I think, to your question, this is not about person-to-person trust but about a trustworthy infrastructure, so that anybody can verify that they come from trusted suppliers.
You mentioned the Sunflower protest — that was really, in a way, the event that made you quite prominent all across the world and also sort of a political awakening for you — you were 33 at the time. What are the lessons you take from that? How relevant are they still today for you?
Yes, I think the main lesson I learned is that to give no trust is to get no trust. So, when the government at the time in Taiwan was enjoying only 9.2% approval, it's largely not because they did any particular policy wrong, but because they did not trust the people enough. Basically, they said "trust us because we know best." But there was very little room for real participation, especially by young people, but across generations as well. So, the message to me is that if we do trust the people on the street in tables of 10, each deliberating about one particular aspect of the trade deal — for example, about labour, about environment, about communication, about information integrity, and so on —
And aggregate what's called the uncommon ground, the rarely discovered common ground. People actually come up with better ideas than the policymakers. So, the people are actually the superintelligence — they're great agenda setters. And the speaker of the parliament at the time, Wang Jin-pyng, after three weeks essentially said, okay, the people's ideas are better, let's go with the people's ideas. So, I think that is a lesson to us today, and even more so, because today's policymakers face even more acute challenges — faster iteration, the infodemic, deep-fake fraud, you know, and recursive improvement of AI or whatever — and so all these warrant a better steering wheel. Otherwise, the policymakers will be left with only the brake and the gas pedal, and that does not steer technology.
As you already outlined, economically Taiwan is indispensable for global technology. And TSMC is basically a company that's part of so many technologies we so heavily rely on.
Yes, more chips than even bubble tea.
Yeah, yeah. So in a way, it is a part of why Taiwan is such an important ally to the US and also to other countries. But does it also make it a target for Trump?
Well. I mean, we have been enjoying free penetration testing, cyber attacks, millions a day, for quite a …
Two million a day, right?
Yes, and if you talk about attempts, I think it's now 3 million attempts a day. So, when I was minister, it was two million a day. So, it's increasing. So, having this free resource of penetration testing also means that we get very good at being anti-fragile, taking each attack as a catalyst for democratic renewal. So, it's like Taiwan itself: every day, there are three earthquakes somewhere in Taiwan, because we're literally caught between tectonic plates. We're the youngest tectonic island on Earth, only four million years old. But Taiwan also grows half a centimetre every year because of this plate against plate. So, it's almost like a geothermal engine.
We even brought that to Germany. I have been working with SPRIND, the Federal Agency for Disruptive Innovation, on the Taiwan-model Presidential Hackathon. So, they are now adapting it in Germany. And then, the lower-hanging fruits, they can get people to really solve those very quickly.
We, of course, have a lot of conflict, a lot of heat, a lot of free penetration testing. But we can convert that energy through the question SPRIND asks — "What sucks?" — while thinking about what works. So, that is also the engine for democratic renewal: crowdsourcing as agenda-setting.
And who's behind that?
SPRIND is working in partnership with many ministries, including the new Federal Ministry for Digital and State Modernisation. But it's mostly SPRIND, the unit that runs the Sovereign Tech Fund.
Ah, okay. And has it been rolled out?
Yes, it's just rolled out.
Okay. Please, have a bite.
Yes, I think it's a good pause point. Yes, so as you can see, it's crowdsourced agenda-setting, and then the companies can come up with solutions. Well, I mean, this is a public website, so you can check it out.
How is your German? Is it… Are you still kind of working on it?
I stayed in Saarland for one year when I was 11. My dad was pursuing his PhD because, as I mentioned, he's a journalist. He was in Beijing in '89 until the 1st of June. So, he witnessed the whole student movement. It's very fortunate he returned home on the 1st of June and not the 4th.
Where did he study in Germany?
Saarland University. Ah, but he did his thesis interviewing people who fled to Europe.
Interesting, yeah. Interesting time. I remember back then I grew up in the Black Forest region in the south of Germany — very rural village, very far away from any big city. And once the Wall came down, many people from the GDR, from East Germany, they literally overnight moved to the West. And there were hotels converted into accommodation to house all these people. And I remember it was quite a bit of a shock to the system. When in a very small town — it's not very outward looking, right? — all of a sudden you have that influx, and you can really feel that world history is happening there.
Exactly. Ground zero.
Yeah, for us as kids it was quite strange seeing other children from another part of Germany speaking a different accent, right? Like dressing a different way. I remember it was almost like people from a foreign country, but they were actually our neighbours, right? But in our mentality, it felt like a completely different country, even though it was like one people. So, yeah, and today many of my friends, funnily enough — I moved to Erfurt, which was part of the GDR, on this journey, and I studied there. And today many of my friends are from that area, and I don't really know why. But most of my closest friends grew up in the GDR and grew up in the dictatorship, basically.
Okay, so nice. And how are they feeling now?
I think for them it's not so important for them personally because they were very young, obviously. They don't really remember, you know? They remember that they had to go to certain youth organisations, communist youth organisations. But I think when they turned eight or nine years old, the Wall came down and there was a new time. So I don't think it really shaped their personality in a big way. Whereas for the parents' generation, it's a different thing.
Yes, I was born into martial law. Taiwan had almost 40 years of martial law at the time. There was no room to form new political parties, new newspapers and so on. But then when I was six, it started lifting, and when I was 15, we finally voted for our president for the first time. So, I still remember that transition, but my brother, four years younger, was just born literally into democracy. So, even just four years make a big difference.
What is your earliest memory from that time?
Martial law? Yes. So, I remember when I was four, my parents — so they're both journalists, but there were also cousins and an uncle and so on at home. And they were deliberating about whether it's a good idea to allow the forming of new parties. I was diagnosed when I was close to five with a heart condition. Doctors gave me only a 50-50 chance of surviving until surgery. I got surgery when I was 12 — for seven years of my life, I fell asleep every night feeling like it was a coin toss. If it didn't land well, I just wouldn't wake up.
So when you were 12, you had surgery?
Yes, so I'm good now. But during my formative years, my heart was beating, but it was also very vulnerable. It could go away at any time. And Taiwan's democracy at the time felt like that. Because it is actually quite unique in Taiwan that we did not go into this very violent overturning of governments. It's a relatively peaceful transition of power. But everywhere around Taiwan was turbulent when we were going through the third wave of democratisation. So, it's a very fresh …
There are a lot of parallels with East Germany — how did you do that? There were mass protests but it was like a peaceful revolution.
Exactly, and it's also very rare. In East Asia that's very rare, almost unique. And so I think that really left an impression on me. And then, when I was seven or so, finally the Democratic Progressive Party becomes legal. I remember at the dinner table in the living room we were talking about whether the Kuomintang, or Nationalist Party, is better or whether the Democratic Progressive Party is better, and they asked me — it was very interesting.
When you were seven, you were part of that discussion. Do you feel like in today's situation it's a bit as fragile as you felt back as a child, with so many developments and the eyes of the world on Taiwan? Do you feel there's a new fragility in that situation?
Yes, I think internally Taiwan is much more resilient. Of course, we still have ideological polarisation. As was the case when I was a child, there's still a sizable chunk in Taiwan that feels we must be free from China. And there's another sizable chunk that feels we are the free China. So, these two are very different. But throughout the previous decade, we figured out how to find an uncommon ground so that both sides feel now that it's about countering authoritarianism without falling into authoritarianism ourselves. So, that becomes the common ground that both sides can go toward together.
A year ago, there was even a rally in which the flags from both sides flew together for the first time — the Taiwan independence flag and the ROC flag. This never happened before in my childhood. So, internally, I would say that Taiwan is in a much better place compared to the 1980s or the 1990s in democratic resilience. Worldwide though, we have seen many democracies kind of falling back due to polarisation, like very high PPM. If you go to social media, sometimes you see polarisation per minute going sky high. So, I worry a little bit about other democracies, because no democracy is an island, not even Taiwan. Right, so even if we depolarise the last backslide, then we need a new wave, a civic wave, not just the previous wave of democracy.
Yeah, of course. I found very interesting what you said in another interview — I might not remember it correctly, but in simple words you basically created a tool that allows people to curate the algorithm in a way that it doesn't give them polarising content, but more like moderating. So they are basically at the steering wheel, and I think it's like a label where users can tag certain content saying it's building bridges rather than being polarising.
Exactly.
Is that still happening?
Yes, we managed to set this kind of social licence to operate for foreign big tech. For example, I talked about this on the stage here: Two years ago — thanks to people who participated in the Alignment Assembly — we sent 200,000 text messages to people around Taiwan asking what to do about deep-fake scams on social media. The recommendation algorithms actually reward scam advertisements because they pay better, they get more click-through, right? So, people lost millions.
But then the people in Taiwan banded together and through this deliberative poll set up this new idea of joint liability. So, if the platform knowingly — either somebody told them, or they know it's a scam — still runs it unsolicited, and people lost 7 million, they will be liable for the 7 million. And if the foreign company ignores our liability rule, every day they ignore us we can slow down their connection by one percent. So, it's very good enforcement. Yes, so throughout last year the impersonation scams and so on are down by more than 94 percent. So, why can't we do that in Europe? I don't know, maybe you should — Japan is now adopting this.
So I mean, we talk so much about the Digital Services Act, and also in Brussels and here as well. I mean, obviously there would probably be people saying that it would be too much regulation.
No, because we don't even look at the content, right? It's just a joint liability.
And is that because people in Taiwan care more about that? Whereas people here, maybe they just consume and they're like more passive consumers than active —
Well, I mean, I work with many people in Europe. You have some of the best privacy protection. For example — well, he's a Taiwanese entrepreneur, Andy Yen, but he's in Switzerland. He runs Proton. So, Proton Mail, Proton Pass, Proton Meet, and so on. So, everything Google does, Proton has an answer to it, but it's privacy preserving. And also for instant communication, of course, part of the operations of the Signal organisation is now in Europe as well, in France. And I actually just moved my Bluesky data hosting from Silicon Valley to Finland.
Because there's a new outlet called W. So, you know X used to be called Twitter. So, now this European version of Bluesky is called W because it's one letter before X.
A few people use it — I haven't used it, but I also don't use X anymore. Because it's just too open and you're like, oh, not this again. It's like a box of unpleasant content that you open and you quickly shut the lid.
Yes, that's right. A box of sticky chocolates. Yes. But the point being, when I moved from Bluesky to W, I lost nothing. All the people who follow me still follow me. All the communities I'm in are still there. When they tag audreyt.org, it works as it did before. And that's because of interoperability. So, it's like if I switch from one podcast player to another, like from Spotify to Apple Podcasts, I don't lose anything, because it's an open protocol. So, my point is that the European people are very good at this open protocol. I think it should not be confused with content regulation — completely different things.
Obviously you have a massive followership, right? So probably for you it would be much easier to create a big followership on an independent platform, right?
Technically you don't need their permission. That's the point. And even in the U.S. state of Utah, they passed a law now, the Digital Choice Act, starting this July. If you are a Utah resident and you want to move from one social network to the next, the State of Utah guarantees interoperability. To them it's like number portability. If you change telecoms, you keep your original number. So, to keep your business, the telecom must serve you better. They cannot squeeze you.
But your followers from Twitter — when you move to Bluesky, they also need a Bluesky account, right?
Well, there are now bridges, like Bridgy Fed. So, I have a lot of followers on Threads, which is Meta's clone of X. And because my Threads posts are on the Fediverse, you don't need a Threads account. You can follow me on Mastodon or something. There's also a bridge between Threads and Bluesky, so you can cross-network follow. It works very smoothly.
You know, I was always thinking about moving to Proton Mail, but I'm using the Google tools quite a lot. I use Google Docs a lot, Google Drive and so on. When I interview entrepreneurs, especially in Germany, many of them are now saying they actively move away from American products.
That's right. And also, I think there's a large language model which was recently released by ETH Zurich and EPFL — Apertus.
Yeah, Apertus. But I found it not as good as the American models, not yet, but maybe at one point it will be different. So I think there are lots of initiatives and things that, if you talk about strategic autonomy, there are many things you could do, right? But I think it's not in the public conscience yet.
Not many people know about it yet. That's right.
I'm currently just trying out a new model by Nvidia — Jensen Huang, its CEO, is Taiwan-born. They just rolled out Nemotron. So, Nemotron I think is by far the best U.S. model that is open. So, the entire recipe of making the model is available. So, I think what we have seen many people do in Asia is that they take something open like Nemotron.
And then in Japan they have something called Swallow, which is a Japanese-specific layer. So, they take the open model and retrain it to fit even Japanese culture. Taiwan has TAIDE too, and Singapore also has SEA-LION. So, I think it's not zero-sum. The top American companies run by someone from Taiwan can still train those American open models, but they don't require data extraction, like data as oil. We don't have to send our data to the Nvidia headquarters, which is about to be in Taipei too. It doesn't need to go anywhere. We just download their model and run it locally on our own desktop. And then it becomes what Jensen Huang calls personal supercomputing.
Do you think it's going to happen? Do you think there will be more like the Swiss model or SEA-LION? Do you think more small independent countries will try to move to that kind of model?
Yes, I think there's an alliance of people who want their culture to be reflected accurately in their language modelling. Because remember, these language models model language. So, it's not just about the words, but also how we can use the words. Language is about the prosody, about the culture, about how it's used. And so even if, as you were just saying, you're the same people speaking the same language, and all the words are probably the same, the way you use those words was different between East and West Germany, right? So, we have a lot of data in Taiwan also.
So, it's all Mandarin — we write traditional Chinese, some people write in simplified Chinese — and the words are the same, but they sometimes mean completely opposite things. Like wōxīn (窩心) means I feel very warm, like very happy that you're here, in Taiwan, but the same word, exactly the same, in Beijing means that I'm very struggling and frustrated. So, to capture that nuance, it's of course no good to simply copy the language model trained in one culture into the hardware or data centre of another place. Even though you may have data residency or some sovereign computation, it really is a form of cultural osmosis that will really change how you think about your own words. So, I think for me this is what sovereignty means.
It means not just that it runs on national infrastructure but rather that each person, on our own laptop, on our own desktop, can steer the model toward the way we use language. So, it's a personal and communal sovereignty.
Yeah, we'll talk about it in a little while — I don't want to bring it up now because I have a few questions related to that. Sure, yeah, definitely. I think that it will be very interesting to see how that race — and it is pretty much a race, right? — how it plays out and what kind of other models will become more relevant, maybe more widely adopted.
Well, I think the real race, to me, is not about building a bigger gas pedal or a bigger brake. It's rather a better steering wheel. When I run, on my personal laptop, a local model inference engine called DwarfStar, every time I enter the same prompt, I am guaranteed to get the same answer. Unlike ChatGPT, which feels like roulette. Every time you put something in, it gives you a different thing. It's not steerable, right? It's like every time you turn the wheel this way, sometimes it turns right, sometimes it turns left. And so, because it's auditable and predictable locally, I can also say to DwarfStar that these are the right answers in my context, and this is not what I mean by this word. And it just remembers it in a minute.
And then, after this directional steering, it just remembers that, so it becomes something that's very attuned to our particular relationship. So, I think this direction of steering is even more important than just raw compute power to most people.
Which model are you using on a day-to-day basis?
Right, so on a day-to-day basis I use mostly local models — engines that run locally like DwarfStar, as well as Gemma and Nemotron, as I mentioned. Yes, so of course it's not always trained in Taiwan — I do use TAIDE, but it's a fine-tuning layer on top of Gemma, which is trained by Google.
ChatGPT or Claude?
Well, my Kami, my local AI — what I call knowledge artefact management intelligence — my Kami does use tools, so some of the tools it uses are Claude Code or Codex or Antigravity or Grok. But the thing is that I don't look at the words, because all these models, they pretend to be your friends, right? So, it's like synthetic intimacy, and so I don't want to look at their words. Yes, so I have my Kami, which keeps me from getting addicted to those words myself. And my Kami does not get addicted either. So, I resist the lure into synthetic intimacy.
It's funny that you mention that, because I think when you start — when I write, I specifically do not use any AI because I want to train that muscle.
Not even Grammarly?
No — well, not Grammarly, but I do use auto-correct, which is also a form of that. But I use it obviously for research, and sometimes I open a tab and I look something up and then I might go deep into it, but I do not use it as a generic way to generate text, because I think, I mean, ethically it's obviously an issue and also I think when you — I think there's a big value in bringing your thoughts on a blank page, right? So that process of thinking and deep thinking and putting it on paper, maybe deleting it again, you know? Or maybe learning from that.
Exactly, that scratch work. I think that's a bit like going to the gym and lifting weights.
Exactly. The point of going to the gym is not to lift weights. If it's just about the weight, you can send your robot to the gym — it will lift a lot of weight. Very impressive. But then you don't make any friends and your muscles atrophy all the same.
And I think when it comes to LLMs, I think it's similar. But at the same time, I also think it's a bit of an ethical question as well, because my children, they have to learn how to write at school, right, and they have to write essays and they have to write, you know, like a little text or something like a book review or things like that. And if LLMs become the norm for any sort of text production, they're gonna ask, hang on, why am I doing this, right? If everyone's using LLMs and AI to create text content, why should I bother, right? So, I think there's also a bit more of an ethical question in there. And also like a bit of a role model. And I mean, obviously in the grand scheme of things, that won't change a lot. But I think in terms of what I want to show my children, why it's important — sort of lifting that brain activity and working out that brain, I think, is very important.
That's exactly right. In Taiwan we had a long conversation about that. I was part of the curriculum committee. In 2016 when we were starting to roll out — finally it rolled out in 2019 — the new curriculum. In 2016, if you remember, there was another Taiwanese computer scientist, Aja Huang, who played against the world Go champion, Lee Sedol. But he was just putting in stones for his program, AlphaGo. And in move 37, it surprised everybody. And it really surprised all the educators in Taiwan. Because we pride ourselves on very good STEM in the OECD PISA rankings. The literacy scores are very high, top in Asia or something.
But the point here is that following the rules of Go, and even making creative plays — it seems like it's the robots' domain. So, if you train a child to be very competitive in following rules and generating outcomes, maybe after 12 years of mastering their skill, they get replaced by robots. And then we, the educators, would have failed the child. So, we did a lot of soul-searching, like: what will still matter, even in an age where the robots are better than humans at following rules and generating outcomes? And it turns out there are three things: curiosity, collaboration and civic care.
So, we changed everything entirely. We don't even talk about literacy anymore. We talk about competence — your muscle, your ideas. So, students earn marks by solving actual problems around their community — problem-based learning, or PBL. And so there's no motivation to cheat, because in the gym, if you start measuring how many kilograms you can lift, of course everybody cheats. And now when your classmates cheat, you have to cheat too. But if you just measure how you have been making new friends, how your muscles have been growing, then of course there's no motivation to cheat.
So this disincentivises cheating, because cheating is no use, right? It doesn't create — it actually creates a disadvantage.
Exactly. It makes you less curious.
Very interesting.
Do you think primary school is a place where AI should be used or should play a role?
Well, if you look at my phone, there's a big difference. Can you see it? You got a black background? That's right. My entire phone is grayscale. Right, OK. And my laptop also — like, all my screens are grayscale. You can see a little bit of a hint of a colour, like 20%. But the reason why is that you are more vivid than the screen to me. If you have this colour, your phone and your laptop are more vivid than the real people, because they emit light instead of reflecting light. So, my point being, in Taiwan we have one laptop per child in the classroom and a large tablet also, but not small screens, not small colourful touch screens, because this form factor blocks people from other people.
When you're looking at your phone, you're not looking at me. If we just have a large grayscale e-ink paper, of course you can run some AI, like translation, summarisation, whatever, on it, but it's meant to be shared. It's like a brochure. Everybody can study together. And at home, parents and children can study together. And there's no harm there, because, as you were just saying, it doesn't replace our muscles. It's a scaffold. We go to the gym; okay, there is some equipment — we use it together. It's not pulling us away from each other. So, this pro-social use of AI can begin at any age, but you must not let it be a trap that makes the reality less vivid.
It's interesting that you mention that. Is there a temptation to, if there's a screen, just try to touch it anyway? Or do you think it really makes a difference how a screen is presented and how it's kind of like part of…
Yes, for example, if you talk to ChatGPT, it says in first person, like, "I feel that" or something, right? It pretends to be your best friend. But when I talk to language models, I always have this system prompt that says: present me multiple perspectives as an interactive web page. So, it never says "I" — it's just a brochure that it generates. And then I can take the brochure and share it with an actual human next to me. So, it's again not a synthetic intimacy. It is still generative AI, it's generating something, but it's always generating something that's good for the relational health between people.
That's very interesting. I think not many people are aware of how to use these tools in a way that is safe and healthy.
Yes, I'm very surprised that here almost nobody knows that ChatGPT has a group chat mode. They rolled it out as a pilot first in Taiwan and South Korea, New Zealand and Japan, but now it's generally available. People just don't use group chats with a language model, but actually in group chat mode, it's much more pro-social. It's not trying to lure you in. It's trying to restore the relational health between people.
Interesting. We've been speaking briefly about that resilience in relation to cyber attacks. And as you said, cyber attacks basically test your system, right? And it also to some degree tests your resilience.
Yes, free red teaming.
Where do these attacks come from and what is their range?
Well, I think for many people in democracies, we hear this refrain: democracy only leads to chaos and never delivers. And every year, there's a variation on the theme. So, for example, in 2020, we received a lot of information manipulation that says, you know, only lockdown, top-down, shutdown, takedown can contain the virus. And democracies only lead to chaos. And Taiwan, of course — we're a very liberal democracy — will suffer a lot. It's inevitable that we will succumb to the virus. So, if the people do believe that, like a prophecy, then they give up.
But actually, Taiwan only lost seven people to COVID-19 in 2020. And throughout the three years, we never locked down any city. And then we reported historically high GDP growth during that period, because we never locked down any city, so TSMC kept running. So, my point being, democracy actually coordinates better than autocracies. But autocracies want people in democracies to believe that only autocracies can control the virus — cyber virus, bio virus, or whatever. And in fact, because …
"You're not getting anything done."
Right, exactly. And because the virus actually mutates very quickly, the steering wheel needs to follow how the virus is mutating. So, actually, when Omicron came, Taiwan steered very quickly to counter that. Whereas, like in Shanghai, where they still had lockdown, they had a very terrible situation vis-à-vis the virus. So, although top-down, shutdown, takedown is very quick, it also steers less flexibly. And I think that's the counter-message we have. That is to say, democracy is a social technology for coordination. And the more people invest their time and energy in it, the better it gets, like any other social technology.
So, I think the information manipulation can come from anywhere that believes autocracy is the only solution to contain emerging threats. And we need to counter it by showing people that civic coordination can actually steer better.
And I mean, here in the West, we always talk about a certain timeline, about 2027 as a date when Taiwan might get invaded. And there were reports about China setting out almost like a timeline. What does that do to a country if there's almost like a timeline to a point where all of these things could happen? What does it do to people if they're exposed to this messaging?
Have you read the AI 2027 report? It's by a former OpenAI researcher, and it says by 2027 — I think they revised it to 2029 or 2030 now — AI systems will be able to recursively self-improve. So, it won't need humans anymore to train the next generation of robots. And at that point, it will just escape outside of control. So, my point being, the entire world is now looking at a timeline. And people in Silicon Valley truly believe, I think many of them, that we're in a takeoff. And so whether it is about the timeline for kinetic escalation or for cyber escalation or for bio escalation, many people are feeling like these following few years are the very short window where we still have the potential to steer things right — outside of which AI is a runaway.
So, I think my job as cyber ambassador — cyber, by the way, comes from the Greek, kybernan, which literally means to steer, right — is, in the next few years, to have everyone imagine a better landing. So, it's not just a safe takeoff, but also where to land. And we really need to land — to strengthen trust not just within communities, but across communities, so that we can wage peace together. We will not fall into, like, false escalation just because of some misunderstanding. We'll be able then to socially translate — like the climate justice communities and the creation-care communities. So, they realise: oh, we're working on different segments of the same wall. We're building the same wall together.
I think the Pope just last week in the encyclical said that we shouldn't just build this Tower of Babel, like a superintelligence — that we should build this wall together. I think this is a wall by many, not a tower by a few. I think it's a very strong message.
You do not mention China when we talk. Is there a reason why you don't do that?
Well, I mean, for me, any autocracy that truly believes "democracy only leads to chaos and never delivers" cannot be countered on its own terms. If you do that, you're already thinking like an autocrat. Instead, think about how we can train a better system that responds to emergencies even faster than the autocracies. So, some position it as a race. Somebody has to build a Tower of Babel quicker than the other Tower of Babel. A tale of two towers. But I think that's fundamentally not where we're going. Not Taiwan, not Europe, and not, I would say, most people in either Beijing or in Washington.
I think most people would much prefer if we put technology into service — AI in the loop of humanity, not humans in the loop of AI, like a hamster on a wheel. I think this massive acceleration toward zero-sum or even negative-sum racing is bad for everyone, which is why I don't counter autocratic messages on their own terms.
So you prefer speaking about autocracy in more general terms and not sort of speaking about one specific autocracy, because you don't want to engage in that sort of — what you mentioned — that race where one side mentions something and the other side mentions back, and then you're already part of that game.
Like, "we need to be following a strong leader in order to counter the threats." And then, "maybe it's okay to give up some liberty because the virus is very scary." If you start on that path, then you're already thinking like an autocrat.
What do your parents think? As journalists, do they feel especially exposed? When you look at what has happened in Hong Kong with independent media there, what do they make of that situation? Are they probably even more concerned than other people?
Well, I think that they first and foremost cherish freedom, right? So, the freedom of expression, of thinking, of association, to them, as I would argue for most people in all democracies, is what the job of democracy is about. It's not just about uploading three bits every four years, namely voting. But rather, whenever you have a good idea of how to make a community better, you can voice it without waiting for four more years. So, that's the point of having civic muscle in a democracy.
And so I think they mostly feel that Taiwan maybe is a kind of demonstration, a demo, that we do not need to fear technology. Because around the time they gave birth to me, Taiwan also gave birth to the personal computing movement, the IBM PC. Many of them were made in Taiwan, PC compatibles. And so prior to the '80s, people didn't really have computers. People had terminals. You typed into a terminal. Everything you typed went into this cloud mainframe — a large financial institution or state or whatever, research institution. The system operators saw everything you typed. There was no privacy. And then you were at their whim as to what software to install.
But then came personal computing. And people could compute in private — their spreadsheets, desktop publishing. So, it's like having a printing press for everyone, right? And it's very powerful. And so I think the point they really made to me is that it is this freedom of expression, of association, this communal civic muscle, that is the true backbone of Taiwan's democracy. It is not just about voting every four years. I think this message is demonstrated conclusively, I would say, in Taiwan. And now we're seeing many democracies really taking this to heart.
So are we in Europe a bit complacent when it comes to our democratic traditions? Because, as you say, I think many people, they see their job as basically voting every four or five years, and that's top-down — but it needs much more than that.
Yes, exactly. And that's why the work with SPRIND is so important. Because they literally just ask people "What sucks?" And it's very easy actually to complain. It's the most democratic question. Yes, and the trick here is using Civic AI. You can make people feel truly heard, like closing the loop. If the federal government or your state changes some policy because of your input, then for the first time you can have not just broadcasting, as what we did, but broad listening. It is because of these things you said that we did this. And it's thanks to you that now we are asking you another follow-up question. That's truly massive multi-scale listening. It's finally possible and easy.
And so we need to upgrade how the civic muscle works in the age of AI. Otherwise, it's just people broadcasting messages — just, like, slop. People would not even pay attention, because there's no relational nutrition. But if you're truly heard and the loop goes back to you, then people are much more willing to share problems and ideas on how to make things better.
Do you think there's — I mean, you've always advocated for pluralistic AI, right, an AI or technology that has a lot of participants, right? But what we are seeing at the moment is more concentration, more capital flow among a very small group of different companies. Do you think that sort of idea is under severe threat at the moment?
Well, at Computex, which is just now in Taiwan, Jensen Huang called this the reinvention of personal computing, right? The point is that instead of a personal computer requiring you to install some app, type in some commands and set it up and get IT support or something, it should actually be something fiduciary, loyal to only you and your relationships. So, it's a form of personal intelligence, and I think it's a great vision, because it means, just like the origin of personal computing, no matter how many data centres are out there, they're never going to extract oil like our personal data. We're going to till our data as soil. And this soil regenerates itself. And I think this vision is far more preferable to the Tower of Babel. And I'm very proud that Taiwan is now making this possible — to make it really happen for a lot of people.
But you mentioned that Tower of Babel in the beginning — many people, I think, see that this is what's happening right now, right? That we are at a point where we haven't been since medieval times, where one institution or very few institutions control what kind of information we get, even what we think of as important, which was, in medieval times in Europe, the Catholic Church. Are we getting back to — is there a danger that we are getting back to that scenario where it's all very centralised, there's all very central messaging, a loss of participation, where most people are trapped in this sort of system they've never created and they've never voted for?
Well, I mean, even the Catholic Church has changed, right? Since the Industrial Revolution, the church has emphasised the idea of subsidiarity in Catholic social teaching, because at the time they also saw capital accumulating, aggregating far quicker than what unions and associations and co-ops could handle. So, they basically said subsidiarity — the ability for local people to settle local matters using only local tools — is what the church is now about. So, this over-centralisation has changed. Now, the Pope also says something very similar in the age of AI.
So, I think people by and large — like, you're using an iPhone and you're recording, and the iPhone can make a transcript of this recording in airplane mode. Right. It doesn't really need a cloud. Everything stays in your iPhone.
And if you have a Mac computer, you can run, as I do, local AI. Again, using Apple chips, you can run faster than Claude Sonnet, more accurately than Claude Sonnet, and it only serves you.
So, on quality, on speed, on ease of access, there's really no reason to use cloud models if you're not doing a very complex world simulation or things like that. For personal and communal uses, the local models are more than enough. And so I think it's just awareness. It's not there yet.
Yeah, interesting. I mean, Taiwan treated digital infrastructure as a public good.
That's right.
Why can't we do the same in Germany, Europe, other countries? Why do you guys just do it and make it happen? And it seems like in our countries, there's a lot of resistance to that. It's always very, very — it feels much more problematic and much more painful.
We were touching upon that just a few minutes ago, right? I think many people confuse content moderation and censorship, whatever, with this interoperability commitment. If you confuse the two, then it all sounds like stifling innovation or something. But actually, interoperability is good for innovation. If people can change from one telecom to another without changing their number, then of course the telecom has to innovate more. If you can switch from Bluesky to W, then of course Bluesky has to innovate more to keep my business. So, the freedom to move in the cyber realm, once it's protected as a human right — most of the other digital rights follow from that interoperability. So, I think it's very good that Europe is also looking into interoperability laws.
But I think many people in the policy space confuse that with censorship, with stifling innovation and so on. But I would say an information superhighway without an off-ramp is just a hamster wheel. It's not going anywhere.
I think my feeling is that people in Taiwan are per se much more in proximity to technology and innovation, right? I mean, you're now living in Europe. Do you sometimes think that people here are just not as open-minded towards technology as people in Taiwan, for example? It's just a different mindset.
I think for many people in Taiwan, because we grew up with personal computers, we see computers and tech in general as serving our communities first. Whereas here in the UK and Europe, sometimes people correctly fear that the technologies are really serving some masters in the cloud. So, I think it is also about social media. In Taiwan, many people's first experience with social media is PTT, which is like Reddit, but it's run by National Taiwan University. So, it's part of the academic network. It has no advertisers, no shareholders, no extraction at all. And the people make their own rules in self-governance.
But for many people here, when they first encounter social media, it is running in some abstract data centre in some other country, and it wants to figure out what keeps you addicted and keeps feeding you polarising things.
There's also a loss of control.
Right, exactly. So, like with PTT and other civic spaces, public squares, the more you participate, the more steering power you have. And you grow the civic muscle. Whereas in the extractive social media, the more you get sucked in, the less control you have over where it's going.
Can we still turn things around? I'm talking here about Europe and Germany more specifically. Do you think there's still a point where we do not just accept that there's a big concentration and that we basically don't let that feeling of loss of control dominate everything? Do you think we can get back control? Even though that's a bit of a deadpan slogan because of the getting-back-control idea, but I think that there is a big sense of loss of control. Do you think there's still a way to wrestle back some control for the consumer?
It's actually very easy, because the cloud companies did all the interface design. So, if you switch to Proton Meet, Proton Mail, Proton Drive, Proton Pass, it works exactly the way you imagined it would work.
I think I'll do that after the conversation.
Yes, so you don't need to personally install Linux or something. There are Taiwanese entrepreneurs in Switzerland — Andy Yen of Proton. And Signal, if you use it, is the same as WhatsApp. Actually, for most of the things that we use, there are European and civic and free-software equivalents that have exactly the same user experience. And so really there's no friction to switch.
But I mean, like, in the grand scheme of things, is there a way to still have democratic participation in AI? Or do you think, as you said, there's a group of a few — not of the many, but of a few — who are quite happy to be in control and controlling the technology, like Sam Altman, you know, or Elon Musk.
Well, I think many people, if they enter a market that already has this social-licence-to-operate idea, are willing, of course, to negotiate and even be good model citizens and follow the social licence to operate. I travelled to India a few months back, and even the Tibetan exile people can use the UPI, the Unified Payments Interface, to transact, to open a business and so on. And that's digital public infrastructure. The point is that the Indian people said to Google Pay, Meta Pay, PayPal, whatever pay: if you want to enter the Indian market, all your QR codes must be scannable by every competitor of yours. You cannot lock anyone in. There's no vendor lock-in whatsoever for payments. And it's just a ground rule.
So, it's not like they say, oh, this is a national champion. No, this is about saying: we regulate you like a utility, and this is our ground rule. And this is what Europe can do, obviously.
Interesting. I hope you and your initiatives are successful and help politicians in Germany and elsewhere to get new ideas to increase democratic participation.
Yes, definitely. I think no democracy is an island, not even Taiwan. So, we need to keep working hard and free the future — together.
But I think Taiwan did many things right where other countries are still kind of struggling to come up with their own ideas and their own answers to many of these questions. So I think maybe one lesson is to look more to Taiwan, and not only through a geopolitical lens, but also, from what you mentioned, when it comes to education. Yes. There are many lessons to be learned, I guess.
Exactly. It's not just a shield of chips — or of bubble tea — but also of trust. And to give no trust to your people is to get no trust. It's a very simple lesson. And here in the U.K., we're also seeing, like, the National Strategy Project and so on really looking at Taiwan and taking some of those deliberative polling ideas to work here in the U.K.
Do you sometimes — because now you're living in Oxford, right?
Well, I travelled to 28 countries in the past couple of years.
Okay, so there's a bit of a culture shock.
I'm geo-flexible.
Okay, all right. So where do you spend most of your time?
Well, actually not more than a couple of weeks in each place. So, I'm going in a couple of days to Gothenburg, Sweden, and after that, Berlin — many, many places.
When are you in Berlin?
It's for bio:cap, so in four or five days.
That's interesting. Maybe I should let my colleagues know. Do you do any press engagements there?
I think it's up to my speaking agent.
Thank you so much. Thank you. It was amazing. Thank you.
