Interview

Interview with Audrey Tang

In 2014, Audrey Tang was a software programmer dedicated to reforming the Taiwanese government as part of the Sunflower Movement. Now Taiwan’s digital minister, for more than two years Tang has introduced digital governance to the country, allowing citizens to crowdsource laws via the online platform vTaiwan. Here, Tang – a self-described “conservative anarchist” – speaks about the future of democracy, radically transparent governments, and fact-checking robots.

The following interview was conducted for Finding Ctrl, our compendium of visions for the future of the internet.

Over your last few years as digital minister, what’s the change that you’re proudest of?

vTaiwan is a project that is recursive. By that, I mean it is both a collaborative space, but it also generates laws and acts that enable more collaborative spaces. It’s a generator. I’m pretty proud of the sandbox system, which is all a result of the vTaiwan processes.

Sandbox – which we actually learned from the UK fintech sandbox – is the idea that you can challenge existing regulations for up to a year, and for a limited time, limited risk, [give] the entire society [time] to get used to this innovation. If it’s not a good idea, it’s open innovation, so everybody learns something.

If it is a good idea, then the regulators, just over 60 days after the end of experiments, merge [the idea] back into the regulation. We moved far beyond only fintech. vTaiwan has enabled the sandbox of autonomous vehicles, of platform economy, 5G spectrum.

Basically, any law, any regulation, can be challenged, aside from money laundering and funding terrorism. Otherwise, you can challenge anything, which is why I think Taiwan is the only jurisdiction to enable multimodal [transport] like hybrid self-driving vehicles.

We have the Taiwan CAR Lab that tests various builds. You can have a flying car or whatever. After testing such builds in public sandboxes, you can then work with the regional revitalisation communities to enable particular use cases, like drone delivery, and autonomous ships for remote islands, and things like that, and challenge – or rather run with a forked version of – the existing laws.1 Tang was instrumental in the hacker group g0v in 2012, which aimed to “fork the government”. Now, g0v continues to work with the vTaiwan project.

1. In software development, “forking” refers to creating another version of an existing software.

It feels as though Taiwanese citizens are very engaged and have a real appetite for transparency. Do you feel like a similar system, transposed to a different nation, would have the same success?

Taiwan is essentially just a larger city. From Taipei to the southernmost high speed rail station, Kaohsiung, it’s less than two hours, so it’s really just a larger city, geographic wise. Population wise, it’s 23 million people, but then it’s easier for us to deliver broadband as a human right, and so on.

I think it’s really a substrate of very high literacy; very high democratic participation; a taste of continuous democracy, not just voting. I think it’s really the substrate upon which this kind of radical transparency governance system is built.

I would say, for other jurisdictions who want to try this, it’s far easier to start at a scale of a metropolis or a city where the situation is comparable to Taiwan, then scale it out and more deeply.

You’ve said in the past that democracy in Taiwan is as old as the World Wide Web. As such, do you feel that other democracies – who are trying to adapt and update governance for the internet age – are less resilient than Taiwan when it comes to online threats against democracy, such as the weaponisation of disinformation?

I think in Taiwan we’re very familiar with disinformation, and also the civil society really takes full responsibility instead of relying on the government.

I think that the core reason [for this] is that while our [first] presidential election was in 1996, the lifting of the martial law was 1987. That leaves a decade where the civil society enjoyed the freedom of assembly, speech, and so on, to build their legitimacy even before the democratically-elected president built some legitimacy of the administration.

That means that for many issues such as disinformation, disaster relief, or things of a public service nature, people tend to trust the social sector more than the public sector. If the largest charity, like Tzu Chi, publish a number about disaster relief and then the government publish a number, chances are that people mostly believe the social sector’s number.

That has been the case since I was a child. I don’t see it changing now. That means that the civil society really feels an obligation in tackling issues like disinformation.

How does this play out in practice?

The g0v initiative, for example, basically takes everything that people think is of public value, but the public service isn’t doing – or shouldn’t be doing, or isn’t doing enough – and they just change the “gov.tw” [link] into “g0v.tw” to introduce the shadow government service.

For disinformation, the Cofacts system is pretty neat. It’s basically a chatbot in an end-to-end encrypted system, which is like WhatsApp.

When people add this robot as their friend, whenever they see a rumour, they can just flag it by forwarding it to the bot, and then the bot does fact-checking through collective intelligence. It’s just like spam mail in a private communication medium. People can also voluntarily report spam or junk mail into the Spamhaus system that’s built by the social sector.

They then partner with the Taiwan Fact Checking Center2 which looks at the most virulent rumours that are spreading, and do a fuller fact check in a way that is very visible, and also in a way that reviews the entire investigative journalism work that’s behind those facts and fact-checking.

This then of course feeds back into the algorithms of popular social media such as Facebook, to dial down their virality so [a rumour] reaches fewer people, which is akin to the Spamhaus sending signals to Gmail and then the incoming email gets moved into junk mail folder.

This is a pretty resilient civil society alert, not at all managed by the government system, that basically lets people see both sides of the story, or many sides of the story. Also, it has a definite benefit in being open source.

The database is open as well, and so third party projects like the Meiyu bot gets built. Meiyu is another bot that people can add to their chat rooms or chat groups of their family channel for example. This bot basically listens to each and every message in every group that they have been invited in, but they don’t keep the log.

It compares it to the Cofact database, so whenever it’s similar, it just says, ‘Oh this is a rumour. It has been disputed, or it has been clarified before.’ People, instead of waiting for five hours or waiting for a day to get a fact check, they actually get a fact check as soon as they post it in their family channel.

This really changed the behaviour of people. It saves intergenerational conflict because people don’t have to correct their family members. Now a robot can do it for them!

2. Launched in 2018 to combat fake news.

And why do you feel it’s so important for software to be free and open-source? What could other governments learn from Taiwan in that regard?

It’s good for sustainability of service. In Taiwan, we have a Government Digital Service Guideline that says ‘open by default’. While open source is important and it is in the guideline, we also say that the open API,3 especially import/export API, as well as the data, the government data that is produced, is equally important if not more important than open source.

I think open API is really the key, because it enables independent service vendors and providers to be in a synergic relationship with existing large vendors. If you don’t have that open API as a kind of anchor to collaborate, then they’re in a kind of a zero sum relationship with existing system integration vendors.

That’s literally the first policy that I enacted as the digital minister, is making open API the national default for procurement. Open data, of course Taiwan is really committed [in that] anything that people can see as part of freedom of information law must also be able for the machines to see as well.

That is because people have different modalities of learning, of understanding, and so on. Some people prefer a bubble graph, some people prefer things that they can interact [with]. Without open data, you’re basically limiting everybody to presentation style that maybe five per cent or three percent of the population are really comfortable with, and that is not really inclusive.

That’s the argument for open data. Finally, for open source, it enables what we call knowledge sharing or cooperation, because then you don’t have to maintain the burden of adapting the system to newer devices, newer requirements any more. You can ask the community to help build it for you, but it is only if you have a community in the first place.

Through open API, open data, the community can be built around public service. Once the community really trusts the government to deliver on its open by default promises, then the open source will find a large number of people who want to co create and maintain with you.

3. An application programming interface that is publicly available. 

As a conservative anarchist, what’s your long-term vision for our networked world?

That depends: how long is a long time? Buckminster Fuller4 has a saying that I really like. He said, ‘You never change things by fighting the existing reality. To change something, build a new model that makes the existing model obsolete.’

Basically, instead of struggling within the rules of the existing system, what the internet has always been about is that showing a voluntary associative system where people can innovate without asking for permission. Actually, it’s a new model that makes all sorts of existing models obsolete.

I think the internet, as long as the core principles (end-to-end innovation, permissionless innovation) are kept, there’s always new generations of innovators that can take their vision of future and imbue it into the code of the internet. I’m not worried about the internet, which is why I’m a conservative anarchist.

Conservative means there is a tradition. The tradition has been going on for quite a while, and what’s important is for people to understand, to embrace, and to respect the tradition of the internet.

Once digital literacy is part of the common curriculum, part of the common sense of a society, then naturally, when people are faced with new threats like disinformation, people will react in a way that adds to the internet instead of taking from it. People will not, for example, react by saying, ‘OK, we ban blocks of IP addresses,’ and so on.

Back when spam was really a problem, during the Bayesian spam filtering days,5 it was really tempting to ban entire countries from sending email, but the spam management community decided against that. [Instead] a supposedly costlier but ultimately more human rights-preserving way of detecting patterns and only blocking systems with those patterns [came about].

Otherwise, we’re taking away fundamental rights – broadband as human right – and the rights to communication to entire blocks in the African continent of people, and that would actually be a shame. I’m very happy that the spam-fighting community made the right choice; I’m making sure that that the counter-disinformation community in Taiwan and also abroad makes the right choice this time.

4. An American architect, inventor, and futurist.
5. An email-filtering technique that often relies on using a multiset of words to identify spam.

Would you say, then, that you are optimistic about the future of the internet?

I’m a possibilist! There are various possible futures. I think I’m optimist up to the point that I think an optimistic vision is self-fulfilling. Then of course, a pessimist, or a doomsday, or authoritarian vision is also self-fulfilling.

We will likely see that the internet being warped in the different thought patterns, different norms of each society, and in some places warped beyond the design of the original internet thinkers, both good and bad. All sort of possibilities will happen – that’s what I’m witnessing.

My favorite singer and poet, Leonard Cohen, said, ‘There is a crack in everything. That’s how the light gets in.’

The internet, to me, is never finished. People keep seeing various cracks or different problems, even in the very core systems like the DNS6 and so on, but then that’s how the light gets in.

That’s why we have a vibrant community caring about this, really a commonwealth of all. I think it is those cracks that keeps the community moving forward and upward as the internet community.

6. Domain Name System.

This interview has been edited and condensed for clarity. The full transcript of the conversation with Audrey Tang is available here. You can find more interviews and ideas in Finding Ctrl, our compendium of visions for the future internet.