COLLECTIVE INTELLIGENCE WITH MORE SIGNAL THAN NOISE
Audrey Tang // Taipei
An Interview with the Digital Minister of Taiwan by Matthew Claudel
Matthew Claudel (MC): Could you start by introducing me to your office, the digital office, and what the mission is?
Audrey Tang (AT): Taiwan is in an ordinary circumstance. We’ve never declared an emergency. There have been zero lockdowns. There are, unfortunately, six people who succumbed to the coronavirus, but six people is not that much by international standards, by far.
Today, our professional baseball league will have a thousand people in the audience. We’ve never had to close schools and so on. We’re under normal circumstances.
This is my office, literally my office. <laughs>
It’s a Social Innovation Lab. Next week, we will be done tearing up all the walls around the office so that it will become a park. Anybody and their dog can walk in and enjoy the vicinity.
The Taiwan Contemporary Culture Lab is home to many artists as well as social innovators who try to communicate how important it is for people to get together to innovate for the public good. That is social innovation.
I’m officially the Digital Minister in charge of Social Innovation, Open Government, and Youth Engagement. My office’s role is to make sure that the best ideas, the brightest ideas, are amplified throughout our network of many social innovation units around Taiwan so that, for example, when people in Tainan think, “Oh, it may be a good idea to visualize the pharmacies’ mask distribution,” then we make sure that all the pharmacies in Taiwan publish their stock level of medical masks every 30 seconds. Now it’s every three minutes because nobody queues to buy mask anymore.
In any case, this real-time participatory ledger enabled everybody to take part in the governance. Instead of trusting the government to make good use of public resources, the government instead trusts the civil society, trusts the entire society to make good use of the shared common data.
So that they can form the data coalition and data collaborative, that make the analysis, that make the best decisions and share those decisions, and also take care of people with blindness who cannot view a map into voice assistance, into chat bots, and things like that.
The main idea is that we should support but not control anything that is a data collaborative by the civil society.
We run an annual Presidential Hackathon, where the top five social innovation teams get a trophy from our President, Tsai Ing-wen, now in her second term in a couple weeks.
The trophy is a micro projector. If you turn on, it projects the image of the president handing you the trophy, promising whatever you did in the past three months will become national policy in the next 12 months. We do shorter hackathons as well, like the cohack.tw, which is specifically around coronavirus mitigation and transitioning to a post-coronavirus world.
Again, we use technology such as Polis, that clusters people’s shared reflections and opinion so we can listen at scale without being swamped by people who AstroTurf or by trolls or things like that. It’s basically a troll-proof way to get people’s rough consensus on any divisive or potentially divisive issues. That’s another thing that we do.
We also respond to the e-petitions. Whenever, wherever people—that includes residents of foreign nationality and people under 18 years old—think of a great idea and mobilize 5,000 people to support them, such as banning plastic straws in the take-out of the national identity drink, the bubble tea.
“The main idea is that we should support but not control anything that is a data collaborative by the civil society.”
MC: You said support but not control. There’s a nuance. You’re not controlling, but you have the credibility with pharmacies that when you turn around and say, “Look, we need you to open the data,” it’s clearly a mutual collaborative effort. I want to understand this dynamic. It’s not that you’re mandating.
AT: Of course. They can stay out of it, but they see that if people go to their pharmacy, they get now, if you’re an adult, 9 masks every two weeks or if you’re a child, 10 masks every two weeks.
There’s more instant gratification if instead of waiting until the end of day to see the statistics, which most citizens never do anyway, after a couple minutes you just refresh your phone and see the stock level deplete by 9 or 10.
This is participatory accountability that establishes trust directly between the pharmacies and the customers instead of through the legitimacy of a government. They are also enjoying a stake in it. They’re national heroes now anyway.
Basically, they get more instant gratification for the feedback that they receive from the citizens, as well as because they are stakeholders in the system, they can suggest new improvements.
We did implement all of that into the data schema that includes, for example, their opening hours, the time that they prefer to hand out numbered cards in exchange for the shorter queue and things like that.
All of that makes sure that their design is also part of the national system, making this again a co-creation. It’s not something that we mandate for the pharmacies. They’re also designers.
MC: I’m curious. Obviously, this, the coronavirus, has given you a very specific set of very clear challenges that your office is very well-suited to. The structure and the accountability networks that your office has, and the digital capacities you have, perhaps, are well-structured to deal with this particular crisis. I’m curious how this moment feels different than the work that you do ordinarily.
AT: The main difference is that the time scale of the crisis is the same or almost the same across the globe. Previously, when we worked on climate change mitigation, for example through our civil IoT network, we work with people who crowdsource their environmental data, measuring air quality or water quality.
It doesn’t have the same sense of urgency when you compare it to the smaller Pacific islands who will suffer the consequence of climate change within one generation. In Taiwan, we might suffer [the effects of climate change] in two or three generations.
If you’re a larger, more continental territory, then you suffer it maybe four generations down the road. The timescale is very different amongst all the different people working on it. That has been the main challenge of working on climate change mitigation.
However, the coronavirus shortened this generation gap into a week-to-week gap. Take any two jurisdictions, they’re only weeks apart when comparing their situation. That expanded our international horizon massively. This bilateral conversation between epicenters has become almost immediate.
“This establishes trust directly between the pharmacies and the customers instead of through the legitimacy of a government.”
MC: I find that interesting, actually, that Taiwan is very plugged into a global knowledge base around this, and that you get this real-time flow of information in a way that hasn’t been as relevant before. Has that been a two-way exchange of information? What are you learning from other places?
AT: Definitely. First of all, I’d like to share this website, which is social innovation. It’s not done by our government. It’s by a bunch of YouTubers, and with all due respect. Taiwancanhelp.us is a timeline of how Taiwan has been dealing with the pandemic, and if you scroll down a little bit, you’ll see a crash course.
This is very important because it’s literally done by the person who wrote the epidemiology textbook still used in universities, but Chen Chien-jen, academician, is also our vice president. <laughs>
MC: Awesome.
AT: It’s very interesting how he is able to then work with professional communicators, to take cutting-edge science and turn it into very easy to access courses that are currently being translated into multiple languages.
Of course, that includes the latest developments of the international scientific community. There’s no way that Taiwan could have done this alone around rapid testing, around vaccination, and around pharmaceutical research.
We also learned from the digital part from, for example, Tokyo, who has offered a real-time dashboard of easy communication web page that everybody can look and see how exactly things are going on around Tokyo, and in Tokyo.
The great thing is that the entire proposal is done by their civil society, the Code for Japan. It’s all open source. I personally contributed on GitHub to take this dashboard and change the language selector to the Taiwan version. Then I worked with the g0v community to basically adapt that not only linguistically, but also culturally, so that people in Taiwan can now enjoy the kind of unofficial civil society port of that technology, and compare things in an apple-to-apple way.
There are many other municipalities in Japan as well as around the world that now use this open source code. I think there are almost 2,000 people forking that repository as a shared way to visualize data.
Our mask distribution data is then adapted by the Korea, who took our idea and convinced their pharmacies and distribution channels to also open their real-time data.
The first working navigable mask availability map in Korea is actually written by Kiang in Tainan, from Taiwan. He doesn’t speak the Korean language, but he knows, I guess, JSON, and GeoJSON, and OpenAPI. <laughs> That’s how international collaboration is done.
It’s not in a kind of traditional track one diplomacy, where you have to first sign a memorandum of understanding (MOU) before doing anything, but rather, people just collaborate on Slack and GitHub very organically. It just so happens that some of the collaborators and researchers are digital ministers or vice presidents. <laughs>
It’s a very interesting way to look at international civil society.
“People just collaborate on Slack and GitHub organically. It just so happens that some of the collaborators and researchers are digital ministers or vice presidents.
MC: I think there’s a question of technological literacy on both levels. Part of what you’re describing in the domestic response has been hinging on technical literacy. That’s everything from pharmacies even having data, to citizens holding government accountable using these platforms, or even just engaging with information and reading graphs.
On an international level, this becomes as rich as it can be because Japan has a phenomenal team working on this, and Korea does, too.
I’m perhaps asking this from the perspective of someone living in the United States, because here, that assumption doesn’t quite hold in many cases, and so I’m curious. Maybe this is embedded in your fast, fair, and fun principles, and the work you have done toward bridging the tech literacy divide.
Can you talk a little bit about that, as a part of your work?
AT: Sure. In Taiwan, we say digital competence. We don’t say literacy. It’s a conscious choice because in Mandarin, literacy or 識讀 assumes that you are a consumer or a viewer, a reader of information, whereas competence 素養 means that you are a producer, a creator, a steward of information, and data, and media.
This is an important distinction because in Taiwan, broadband is human right. No matter how remote you are or how high you are on the peak of Taiwan, almost 4,000 meters, you still have 10 megabits per second at 16 USD per month, with unlimited 4G connection. If you don’t, it’s my fault.
The point is that if the marginal cost is literally zero, because there’s no extra bandwidth charge, there’s no extra charge in equipment for anyone to be a YouTuber, to start their own media.
The kind of digital competence framework that we baked into the basic curriculum, starting from the primary school, from the first grade, actually, is to show people to be responsible citizen journalists, essentially.
The importance of fact-checking, of understanding your sources, of framing correctly, of engaging with the audience in a responsible way, of tapping into the journalists’ community, of how you can, for example, listen to the presidential debate and help typing in the transcript and fact-check each and every word they say, and your name will be credited into the collaborative fact-checking, and things like that.
Basically, all of it is with this civil society mandate, that the social sector points to the direction we should go in, and the public sector takes that direction and implements it by convincing the private sector that this is really a more sustainable and useful way for public benefit.
It reverses the dynamic of traditional public-private partnership, whereas in the US, many times it’s the private sector with their massive R&D and consumer reach telling the government that, "We’ve already reached a pact with our consumers. It’s best if you do some regulatory adaptation or reform that takes care of this new consumer relationship between the consumers and surveillance capitalists,” I’m sorry, “and the capitalists.” <laughs>
In Taiwan, it’s entirely the other way around. It’s the people who have joint control of the data. For example, citizens decided to apply pressure to the government. Our environment ministry had to submit to the legitimacy of the thousands of people measuring their own air quality.
[The government started] working with them to apply social pressure to the private sector industrial parks, so that they have to agree to also start measuring their air quality on the land within the private industrial park but the land is private property, and so on.
This is a reverse of the traditional PPP partnership model. I think this is by far our most fundamental assumption, a social sector-led assumption when it comes to digital social innovation.
MC: Something you’ve brought up a couple of times is legitimacy. I think that’s at the center of all this, because not only does your approach have tremendous legitimacy, clearly, in the public sector, it also has tremendous legitimacy in the broader social structure of Taiwan.
Something about that kind of deep model where you have education and digital competence built from day one, from a young age, that cultivates a condition where you can respond quickly to a situation like coronavirus, because you have a framework of legitimacy.
AT: That’s right, the collective intelligence has more signal than noise, and the government shows that they do respond to all the suggestions… [The minister hosts] a daily live-streamed, “ask me anything” press conference, where he responds to all the digital social innovations yesterday, and applies them immediately. That massively increases mutual trust.
MC: That’s fascinating. It’s something, that nimbleness, that capacity for responding quickly, and for that response to be taken seriously.
Have there been situations where you’ve made a fast response and with good intention, but found that it was the wrong one? How do you manage that communication, where you say, “OK, we tried this. It didn’t work, and that’s part of figuring out the right solution”? Can you talk about that process?
AT: Definitely. For example, in the very beginning of wearing masks, there was no rationing. We just distributed all the stock of medical masks to the major convenience stores, and we said everybody can only buy three pieces each, at a very reasonable cost.
However, there’s no way to work across stores or across different chains and convenience stores, to make sure that the same person doesn’t ride a scooter. If masks arrive by 2:00 AM, they can just ride a scooter and buy all the supplies, three pieces at a time.
If you have sufficient numbers of people doing this, then by 6:00 AM, there are no masks left to purchase.
MC: Did that happen?
AT: Yeah, that actually happened.
MC: Wow.
AT: That was I think the first five days of the mask distribution policy. There was massive unrest, speculation, conspiracy theories. There was no insight into the stock levels of any convenience stores. It created unnecessary tension between the store staff and people, and so on.
It went on for like three days. Finally Howard Wu of Tainan, he was personally seeing too many instant messages in his friends and family channels saying, “Hey, this convenience store still has some masks,” or “Masks are running out at this convenience store.”
It wasn’t very useful because the way they communicated geolocation was not commeasurable. It just added to the noise even more.
He decided to code up a simple map using the Google Places API where all his friends and families can report whether the masks are running out of stock, or there’s plenty of stock, or it’s something in-between. It’s like Ushahidi. It’s like a cross-list intelligence.
However, he didn’t anticipate that everybody around Taiwan who saw his work through the television would start to use his service, and so in just less than a day or so, he owed Google some 20K US dollars, in API usage fees—not really another number. <laughs>
He had to shut that site down and look into alternative ways, but eventually we worked with Google. In an act of, I’m sure, social responsibility, Google waived all his usage fees.
I showed this prototype, which only lived for a couple of days, to our premier, our prime minister, the very next Monday. I said "Look, if we can get accurate data in pharmacies, then this would actually save the mental health of our entire population. Nobody needs to waste their time queuing or going to pharmacies that are already out of stock. They might still queue a little bit at the beginning, but we will see exactly how the supply has been growing.”
In any case, the premier looked at that and said, “Oh, it’s just like navigation software. If you see a road that is shorter but red, you know it’s congested, so you shouldn’t take it. If you see a pharmacy that is further, but it’s green, it means that you should take it despite it being further away.”
It’s a good metaphor, so he just said to the National Health Insurance Agency, as well as the entire ministry of health and welfare that says we should do everything to support this social innovator, and we need to absorb the cost.
Now, the leading map from Kiong, another developer using OpenStreetMap technologies, is running on our National Center of High-Speed Computing, which is, I think, one of the top 20 super computers in the world. They absorb all the usage fees, all the bandwidth fees, and so on.
This piece of civic tech actually became a kind of civil engineering because, like roads and bridges, it’s used by over half of the population. It shifted from the idea of civic tech being used by only a handful of people, into civil engineering, which is used by pretty much everybody.
"Joy travels quicker than outrage online... it is really the only effective way that we found to counter disinformation."
MC: Digital infrastructure, yeah. That’s fantastic. There seems to be this really interesting back and forth relationship between programs that are created by government, and things that are coming up and then sort of how they flip and flop and grow.
AT: Right. The rallying cry of the g0v community, where all these developments of mask pharmacy applications happens, is to “fork the government.” Very peculiar pronunciation, fork the government. <laughs> Which means that instead of writing the government off, you take it into a different direction. *
Whenever people are fed up with a government service, they just create an alternative, changing something.gov.tw into something.g0v.tw. Just a change of letter in your browser bar, you get into the shadow government that is jointly governed by the people.
It is always open source and creative commons content, so when the government sees that, “Oh, it’s a good idea,” we just say, “Oh, we can’t beat them. We join them,” and then we make sure that they run on government-sponsored property and thereby become infrastructure.
MC: I think that’s brilliant. It sounds like you also had that happening with scientific experts in the hackathon, and when people were giving ideas, and there were experts saying, “OK, this is workable. This is not. This conflicts with certain policies…”
AT: Yes, there are also legal counsels for essentially human rights and impact assessment.
MC: Which is of course an important and interesting thing, when you’re creating rapid response to a public health crisis.
AT: Right, because everything that we do needs to be operating under normal law. As I said, there’s no emergency situation, so we’re still operating under a traditionally continental law system, where everything we do needs to have a legal basis as pre-approved by the legislature.
The Minister Chen may now have 93 percent support, but that also means there are 7 percent of people who disagree with his measures, and we thank those 7 percent of people because they keep us honest.
They demand that we explain the scientific reasons, the scientific case of every measure that we do. Without those seven percent of people constantly demanding, for example, “Why 14 days of home quarantine?” we wouldn’t be able to communicate as effectively.
It would be a traditionally top-down health authority thing, but actually, Minister Chen, and Vice Premier Chen, and Vice President Chen—they are all of the Chen family, not immediately related – but in any case, they all take [seriously] this idea of co-learning. Although [experts] may teach or may have learned epidemiology, what they have learned was classical epidemiology.
We’re in a digital world, and digital epidemiology is a chapter that’s currently being written. Nobody knows how exactly digital epidemiology works, but we need to constantly look into the kind of norm that we want to have as a whole society, instead of relying on authoritarian ways.
People above 30 in Taiwan remember not only SARS, but also martial law, and nobody wants to go back there.
MC: That is actually one of my key questions—it seems like a big asset that Taiwan has is the recent learning from SARS. The fact that you went through something similar and you developed some knowledge from that.
First off, what did you learn from SARS that you are now taking into the COVID situation? And secondly, knowing that you were able to learn from SARS, what are you taking forward into the next, into the future, from COVID? What are the provisions you’re packing from this moment?
AT: From SARS, we learned three important things. The first is obvious, because we don’t have an easy-to-cross border. We’re a bunch of islands. It’s impossible to accidentally stumble upon Taiwan.
If you do testing and quarantining at the borders, you don’t have to do Bluetooth tracing apps of any kind. We’re not rolling that out. If you do the quarantine well at the borders, test well at the borders, and act early, like when Dr. Li Wenliang, the whistleblower post that there’s new SARS in Wuhan around the end of December.
We received that information from social media signals on the last day of December. The very next day, the first day of 2020, we started doing health inspection for flights from Wuhan to Taiwan. Ten days later, the WHO would still be saying, “No, you shouldn’t do that. People from Wuhan are OK. There’s no clear evidence of human to human transmission,” and so on.
Basically, because SARS is in the [recent memory of] our society, we took measures as if SARS had happened over again. Nobody knew what this SARS-CoV-2 was like at that point, and we enacted these inspections and soon quarantines at the borders.
The fast response was really important. That’s the first lesson we learned from SARS.
The second lesson we learned is that masks are essential. In Taiwan, masks are very interestingly billed as something that protects the wearer. Of course all of us know it primarily protects other people, but it leads to an interesting social situation where, in a crowd of 20 people, even if only two people wear a medical mask, they gently remind the other 18 people to take care of their own health by putting their mask on.
It’s a very interesting incentive design that was very well-ingrained in Taiwan, and not only in Taiwan, but also Hong Kong and so on, during the SARS epidemic. That norm has been re-invoked during the coronavirus.
Finally, communication is essential. Even before the coronavirus outbreak, we had been using a “humor over rumor” way to counter disinformation—with very fun, memetic pictures. Joy travels quicker than outrage online.
It is really the only effective way that we found to counter disinformation, of course armed with real data, and journalism, and fact-checking, and so on. The payload still needs to be journalistic, but the [deployment] needs to be fun and memetic.
That’s something we also learned from SARS, in that, if you only broadcast public service announcements filled with scientific jargon, it has zero legitimacy compared to the conspiracy theories. <laughs>
The early communication, and a hotline, and a daily press conference, the spokes org of the ministry are all essential designs that we learned from our previous failures in public communication and risk mitigation.
Now, what are we taking forward? I think there are two things that we’re taking forward. The first one is that legitimacy-building is almost an art now. We give out lots and lots of medical masks… the medical masks that we donate to the world are actually dedicated by a lot of people.
Everybody is guaranteed to have three masks per week. If they did not collect some of those rationed masks because they still have some, they can go to the same app where you pre-order the mask, and say, “I want to dedicate my uncollected quota to the international community in need.”
When the recipient of the medical mask [somewhere in the world] receives our gift, they can see that these masks are not a gift from the government of Taiwan only. They are from the people of Taiwan, and not an abstract people of Taiwan, but exactly 500,000 people. And the names are all there. This gives international collaboration a publicly approved mandate.
At the moment, there have been more than four million masks from 527,000 dedicators, and you’ll find people’s names on it. We don’t need 100 percent. There’s people who did not participate in this, of course.
But this is, I think, a social innovation in diplomacy that we’re taking it forward. For international humanitarian aid and things like that, we make sure that this carries the individuals’ contributions, not only their names, but down to the specific amount of contributions that they made. I think this really makes us a republic of citizens. That’s a really new thing.
We’ve had this slogan, Taiwan Can Help, for years, but it’s only after the coronavirus that it actually gained international recognition.
“I think the government should fully trust the citizens”
MC: Right. I’m curious how you see that landing in other places, because Taiwan is obviously, in very logistical sense, Taiwan is really exceptional, right? Because you have essentially a city state, a small population, and one that’s…
AT: 23 million is not a small population. We just have a small geography. We’re very densely packed.
MC: It’s politically different. I’m curious how you see, given any kind of help of the kind that you’re already seeing with some of the digital things, or some of the responses, or whatever might emerge in the future.
Let’s say I’m here in Boston, or I’m the mayor of Amsterdam, or I’m the prime minister of France. How do you see this landing in other places?
AT: Maybe it’s a good opportunity for us to look past the traditional jurisdiction-based multilateral point of view. It’s much more useful if we just compare one epicenter to another and see how they resemble each other, and only [collaborate] on the paths of where they resemble each other.
For example, the Czech Republic uses a lot from the playbook of incentive design that I just mentioned around mask use. They did a very successful, fun digital communication campaign that basically changed the norm around mask use over the course of a week.
What I’m trying to say is that if #MaskForAll unites people who are in similar stages, then we can collaborate on that particular subject with all the subject matter experts.
That’s why Pol.is, the technology infrastructure, creates focus among a swathe of social innovations. It is important in Cohack and similar setups, because otherwise, more people joining actually detracts from the problem-solving, because people have to use their mental bandwidth to look through all the ideas.
For people to cross-moderate each other and then find commonalities between two very different epicenters [makes sense when there are] one or two things—actually, no, it’s seven things or so—to focus on, that is very useful. That’s, again, how digital deliberative technologies can help to unite the research focus of all the resources in various epicenters.
MC: That’s brilliant, and it actually comes into all sorts of different domains, right? Like you could apply that same principle to any number of governance or regulatory questions.
AT: A [misconception] of Taiwan’s design comes from the idea that the social sector had years and decades to build legitimacy. But the president only had comparatively a small amount of time. We had the first direct presidential election in 1996, after the World Web.
Not only are “digital” and “democracy” the same thing in Taiwan, with the same generation of people working on it, but we also need to respect the social sector legitimacy that was already very well-established before even the modern liberal democratic system.
That, I think, is part of the underpinning of why we always say we cannot beat the social sector. We must join the social sector. We cannot control the social sector. We must support the social sector. Even as a language choice, we never say the third sector because they’re the primary sector.
MC: That’s fantastic. Do you have any concluding thoughts or things I missed that have been on your mind, maybe even on a more personal level?
AT: Yes. I think the government should fully trust the citizens. When we say mutual trust, all too often, people think it’s the people who should trust the government, which will lead to people trusting each other.
This interview has been edited for readability. The full text is available here.
Notes:
* Forking refers to making a copy of another author’s open-source code, in order to make changes without affecting the original. It is a way of building upon existing code and contributing to code development.
Author Bio:
Audrey Tang is Taiwan’s Digital Minister in charge of Social Innovation. Audrey is known for revitalizing the computer languages Perl and Haskell, as well as building the online spreadsheet system EtherCalc in collaboration with Dan Bricklin.
In the public sector, Audrey served on Taiwan national development council’s open data committee and K-12 curriculum committee; and led the country’s first e-Rulemaking project. In the private sector, Audrey worked as a consultant with Apple on computational linguistics, with Oxford University Press on crowd lexicography, and with Socialtext on social interaction design. In the social sector, Audrey actively contributes to g0v (“gov zero”), a vibrant community focusing on creating tools for the civil society, with the call to “fork the government.”