Thanks again for making time. I know you’re quite busy, so I really appreciate it.

Mm-hmm.

Really quick, I just have a few basic demographic questions to ask you for the project. It’s just standard protocol and then we can get on to more interesting stuff pretty quick. The first one is, what city or village do you live in?

Taipei City.

Of what country are you a citizen?

This is a very tricky political question.

Yeah, I guess particularly in your situation.

I’ll refrain to answer.

Awesome. That’s OK. What is your ethnicity? You can also refrain from answering if you want.

I think I’m, well, ethnic Han.

What languages can you speak, read, or write in?

Speak: Mandarin, Taiwanese Hoklo, English, a little bit of German. That’s pretty much it.

What would you say is your native language?

Perl.

Sorry. Mandarin. I first learned Taiwanese Hoklo but I’m now more fluent in Mandarin.

[laughs] That’s a new, unique answer. Have you traveled or lived outside of your country?

Yes.

Where did you live?

Germany, near Saarbrücken in Saarland.

Yeah. I was in Freiburg for a year.

Oh, cool.

Could have been farther away.

Right. [laughs] But otherwise, no permanent stay for more than six months at a time.

What was the purpose of your stay? Was it personal, business, academic, tourism?

It was personal. I was 11. My dad was pursuing his PhD there, so I went there for a year.

How many years of formal schooling have you had?

That’s a trick question.

I dropped out of junior high when I was 14. If you don’t count kindergarten, then it’s about seven years. But then I went to grad school classes right after dropping out, so I don’t know how to count that. There’s no diploma, so I guess it’s seven years.

What is your highest level of education?

I dropped out at junior high school.

That’s it on the boring stuff. Thank you for your patience. I’ll give you a quick intro to the project. I know I talked a little bit about it in the email, but just so you know a little bit more about what we’re doing.

It’s called COMPROP, which is short for Computational Propaganda. It’s a joint project between the University of Washington and the Oxford Internet Institute. Broadly, what we’re interested in is any use of computational propaganda, where the digital sphere and politics mix.

What we’re really focused on with the project is automation with propaganda and specifically, bots. That’s what we’re most interested in, but that’s not always present. If there’s broadly any digital or computational propaganda out there, we’re also interested in that.

Unlike other people I’ve talked to, obviously you have a formal position in the government and you have lots of experience with this kind of stuff. If anything’s too sensitive to talk about or you’d rather not talk about it, feel free to let me know and we can move on.

That’s the outline of the project. I guess I’ll start with asking you what a normal workday is like and what your responsibilities as general minister are, just to give me a better idea on your background.

If it’s a Monday, a Tuesday, or a Thursday I would wake up at 8:00, then go to the administration building at around 8:30, and then start meeting after meetings. Then go back to my dormitory around 5:00 to 6:00 PM. Then, actually starting to do real work, which can go to 9:00 or 10:00 PM or so. That’s a typical boring weekday, Monday, Tuesday, Thursday.

Wednesday and Friday are my teleworking days, so I don’t go to the administration building unless there’s emergency meetings. I try to keep them to the minimum. I still wake up kind of early, but I do a lot of analysis, coding, just hackathon stuff.

Every Wednesday we have a hackathon in the afternoon with the civil society and private sector people. Every Friday we have a hackathon with the participation officers in all the ministries. That’s the community-building ideation and everything. That’s Wednesday and Friday. Yeah, Saturday and Sunday are not workdays, so I don’t have to answer that question.

Then my responsibility, broadly speaking, is implementing the open government plan, which entails radical transparency, civic participation and building accountability trail of as many policy-making as possible.

Also diversity inclusion plans, which means that people of all kinds of cognitive modes into what has been roughly the equal way, not just people who specialize in code, in text or in law, for that matter. That’s pretty much my mandate.

I’m really excited about your radical transparency stuff and all the work you’ve been doing in the government. I found about you reading broadly about digital politics in Taiwan. Everything I’ve read and everything I watched has been really impressive. Thank you for the good work you do just as a citizen of a nation that doesn’t happen to be Taiwan, I still appreciate what you’re doing.

This is all really exciting stuff. Obviously one of the main things I want to dive into, which you may or may not have a lot of info on, is bots. The idea of automated scripts’ being used for potentially propagandic purposes or political gain. Not necessarily on the governmental end, but by any political force within Taiwan.

I don’t know if you’ve heard much about it, but it’s a phenomenon that’s been growing a lot in the last few years. In Russia, they’ve been used a lot for obviously misinformation. The Boris Nemtsov murder was a big hot spot. They’ve been used to flood hashtags in Syria and Mexico.

But there have also been benevolent uses to promote constructive debate, refuting anti-vaxxers and anti-climate change people online. I was just curious if you have seen any evidence or had any hunch that that type of stuff might be going on in Taiwan.

That’s a overly broad question. Could you be more specific...?

Yeah, for sure. Obviously bots, especially within the digital sphere, is a really broad term. Political bots are kind of what we’re interested in. More than that, I guess from people I’ve talked to and the research I’ve done, my suspicion is that they may be being used for campaigns or maybe for 網軍 or something like that, on either side of the debate.

Because bots is, broadly speaking, anything that automates human work, right? You mean bots, roughly speaking, in social media...

Definitely.

...on the same public sphere.

Yeah.

Which masquerades, to a certain degree, as a normal human user. At least it uses the same mechanism the human users use to interact. Is that the idea?

Yeah, definitely, on social media.

For an automated program that lets me find out, for example, like Google Alert, if my name is being mentioned on the media. Then it lets me know about it. Does that qualify as a bot in your study?

Yeah, I think that’s a little more tangential, but I’m interested to hear anything that intersects that you think would be of interest. That could be one.

Right. Because it’s pretty easy at this moment to tell a bot’s behavior, versus a human behavior, given a sufficient long observation time. That is to state your intentions not being conclusively asked. Maybe in a few years’ time that will change.

For me once we pass the Turing test in a convincing way most of the time, then there is much more use for people-masquerading bots to appear. Whereas at this moment, if you want to be really effective, I think putting a lot of effort in making bots appear like humans may be actually counterproductive, and augmenting tools like batch automation of querying, of replying, maintaining FAQ, or anything like that.

That is to say doesn’t try to masquerade as human but try to augment humans’ online behavior is much more efficient both in my own experience, and in my observation of use of social media here in Taiwan.

That’s definitely true. One thing that we’ve seen within the project is that bots have benefited from the Twittersphere, in the sense that linguistically and interactively, it’s quite limited. With that limited amount of interaction that you get, it’s a little easier to pass the Turing test and convince people that something may be real when it’s not.

One thing that’s interesting about Taiwan is that the social media scene is different from the general West, and it’s different from Mainland China. As far as I understand, Twitter isn’t super popular.

It isn’t used much. It’s like tenth place or something.

Yeah, you got Facebook and LINE at the top, and then as far as...

And PTT.

PTT as well. I thought to talk to you a bit about PTT, and whether or not you think there might be some sort of political propaganda going on there.

Of course. It’s the main. It’s the core. Facebook is the peripheral.

[laughs] Are you talking in terms of social media usage in general, PTT is the top for Taiwan?

I wouldn’t say it’s the top used, but we see a lot of propaganda starts as PTT-originated content and repost out to Facebook and to mainstream media. PTT start a very active net. It’s a little bit like Reddit in this sense, in that you can very quickly gauge your message’s reaction on different kind of audiences because there are different subreddits.

That is to say, the discussion forums in PTT already comprises of people who already self-identify as some kind of sub-continent in the memetic sphere. People engage in PTT discussions deliberately to try to formulate some meme that they know will resonate with this core group. It makes further spreading much easier.

You can’t do that as easily on Line or Facebook.

That’s a really great answer. What kinds of things have you seen or do you suspect are going on on PTT? From what you’ve said, and what I’ve read as well, it seems like manual propaganda is the way it’s going.

Like Wing Chun, humans basically spreading political messages, stuff like that. Do you...?

PTT is text-based. It’s by far the easiest according to your categorization than the most limited form of expression. You can’t just ask the person you talk about on PTT to reveal anything about their real identity, start a live broadcast, or anything like that.

It’s pseudonymous. It’s a mix between 4chan and Reddit in some places. It’s normal to be pseudonymous. As I said, I think there are accounts in PTT that are semiautomatic controlled. They are bots in the sense that if you post something, they will automatically upload it, download it, or something like that.

Basically, anything that you have seen that’s done in Reddit, maybe there’s its counterpart on PTT. It’s pretty well known here, actually. It’s not a secret or anything.

In that same vein, could you characterize political discussion on PTT as exactly the same as you’d see on Reddit or what kinds of things you talk about?

Yeah. There’s forums dedicated to political discussion, like the ironically named HatePolitics board, which comprises of people who hate politics, but that’s all they talk about. I guess it’s an identity too.

There’s also much more uplifting public issue boards. There’s the PublicServan board, who are career public servants, actually. Then there’s of course always the gossiping board, which is gossiping.

A lot of those propagandas start as so-called 爆卦, which means original discovery, original gossip. I would say, by far, the original gossips are the sources as much of the political discussions.

For example, the quarter-million demonstration back in August, 2013 was started on PTT by a 洪仲丘 case about a military misconduct, a lack of trial, and lack of transparency information of Hung Chung-chiu’s service, an "accidental murder" in the army, and became a quarter million people protest by the sheer amount of 推文 and up voting on PTT.

The original organizers, who didn’t know each other, are all PTT users who eventually met in a café somewhere, but most of the organization on PTT without knowing each other before. It’s not the exception, it’s a norm that there is a lot of these civil society activities that just starts on PTT as a 爆卦.

Just to be clear, it was on the 八卦 board. It based in fact, or was it based on a rumor, the quarter-million demonstration?

It’s the same with gossiping board, is that there’s no need to prove one way or another. The exception is, of course, people would talk around a social object like a news article and provide their own 卦點, which means other things that the reporter did not write. There’s also this critique reading part of the culture in the gossiping board.

For original gossips, no, there’s no need to prove anything. Of course, people would try to come up with pictures, or with videos, and so on, but if they don’t, it’s considered a norm. The moderators won’t do anything.

For that specific example, was it a hullabaloo about something that wasn’t real, this accidental murder you had mentioned, or were there protests that were based in fact about a real situation?

It’s a bit of both. The main protest was about the lack of transparency because it was a martial court. It happened as a misconduct in the army, so there was a martial court. CCTV sent their related information by the much more confidential martial court law. There’s no need to go through the normal court procedure, things like that.

I think there’s a lot of rumors, but there’s a factual basis, which is there is no clarifications and no room for clarifications and for due process if this kind of case is being court martialed instead of tried on the civil court. Eventually, the protesters solidified their ideas, and then finally demanded that it means that because it’s not wartime, it must be tried in the civil court.

That became their main aim after a lot of back and forth. Because half a million people went to the street, there’s tremendous social support. Eventually, that demand is met, and cases like these are now tried in the civil court. So-called original discoveries that was posted during that time are not altered, but they all served as propaganda.

Thank you. That’s fascinating. Thank you for bringing that up. One thing that related that I would like to dive a bit deeper on, the idea of 網軍 in Taiwan. I did read a paper by a professor at Táidà. His name is Chen Shishi. He wrote a paper about PTT and the 2014 mayoral election in Taipei, which I’m sure you know quite a bit about.

Essentially, they found 19 accounts that the suspected were 網軍, and they were posting messages, posting articles that either displayed the Sean Lien in good light, or talked bad about Ko Wen-je. There was this propagandistic thing going on with that election.

I’m familiar with that research.

Do you think that kind of phenomenon, basically 網軍 in Taiwan are quite common, either in PTT or in other social media?

Yes.

How would you characterize how widespread it is if you had to guess? I’ve talked to people who have claimed that they know people who are paid to propagandize for a certain candidate within Taiwan. Is that something that most people think is going on?

Sure, but even unpaid 網軍 is commonplace. It’s difficult to characterize the nature of the practice, because if you say only people who use automated tools are 網軍, then anyone who even install a plugin is 網軍, right?

On the other hand, if you classify it based on it must be politically motivated and paid by a certain politician to further their purpose, of course that reduces the amount of people. As I said, it doesn’t have to be a lot of people at the core.

All they have to do is crafting some memetic device, and then the viral nature of social media will take care of the rest. The people who then spread those news are incidentally 網軍, but they are largely unpaid.

There’s the useful idiot situation, which is a dysphemism for it, about people who are joining in on the mobbing.

One question I’d like to ask is, within Chinese, what normal people understand under the word 網軍. Is it someone sitting behind a computer and posting their opinion freely, or is it people who are paid by politicians expressly to spread a specific message?

Before the 2014 election, it’s mostly used as cyber-arming, which means people who disrupt security infrastructures and so on, which is much more serious. It’s like the cyber-arming kind of stuff.

During 2014, because, as you mentioned, Sean Lien and Ko Wen-je both accused each other of additional propaganda, they brought up this term in 網軍. Then afterwards, it lose any meaning whatsoever. Normal astroturfing are sometimes described as 網軍 as well.

This term currently doesn’t have any definite meaning.

That’s very helpful. In the research, it seems like it might not entirely mean what it means in English, so it’s good to hear...

The idea of one, if look at Wikipedia’s 網軍 term, they actually redirects to 網路特工. That means government hired. There is also 婉君, which I don’t even know how state this in English.

It sounds like 網軍, but it also sounds like a person’s name. 婉君 is used sometime mockingly, and Wikipedia list astroturf as the first use of 婉君, but they’re clearly not politically motivated. They’re marketing people.

We could say that originally, 網軍 had some meaning, and as I said, a security state style of hacker kind of meaning. But then because of its use in election, subsequently lose any meaning whatsoever, and then people just say 網軍 to mean anything, in a way.

That helps quite a bit. You talked a little bit just now about how this term changed during and after the 2014 election. That ties into something I was interested in asking about anyway, which is, how do you think the 2014 mayoral campaign impacted politics and the perception of tech and politics together, both from a politician and the citizen standpoint, if you can answer that.

I would say that the 2014, and the Sunflower Movement before that and 洪仲丘 Movement before that, demonstrates that the speed of communication and transmission is based on the idea of outrage, motivates people to communicate and form communication structures that are far more effective, at least, regarding their issue than anything we have seen before.

Not just through social media, but also through ad hoc programmed networks that connects diverse groups together using a collaboration, rough consensus-like protocol. That’s been studied extensively, so I will not repeat the results.

The instrumental use at this became a lot of interest, both on the actual tools, like Hackfoldr and so on, that was used during the Sunflower Movement, and also the kind of organization, the multi-centered organization, the people want to start studying or even emulate.

As a political worker, what I’ve been doing is introducing career public servants to this kind of self-organization and horizontal communication structures. The first thing I did as digital minister, literally, is to recompile the Linux kernel.

One of kernel machines so that I can install Sandstorm and Docker clusters, and introduce the open-source collaboartion tools, all sorts of collaboration tools into the intranet so that people can collaborate in a way that they know are protected by our cybersecurity team, but nevertheless regain some organizational features.

These kind of self-organized mobilization groups across departmental breaking out of silo kind of ways. That’s why we do weekly hackathons. I was just back from one. It was a very lively discussion.

I don’t think there’s necessarily a difference between my role as a political worker, and my role as a citizen. For me, since I work on free software, free speech, and whatever movements, the original blue ribbon campaign. That was 1995 or something. It was always the same idea of promoting autonomy.

For me, there’s no real difference. I think the main difference is to base participation is not much higher. You don’t have to own expensive computer, modem, or anything like that, anyone can just use this piece of glass to participate, expanse the base.

The basic ideas are more or less the same. It’s just this newer organizational structure and the technology that informs it, which then shapes the generation of organization, which then shapes the next-generation.

Technology is what they call a recursive public or something like that.

Have you found that your plan to reorganize the Linux kernel and do this horizontal communication structure has had tangible benefits so far?

Yeah, it’s very useful. We got buy-in from all the ministries which all recommended one to three people as administration officer, and there is some real policies already being made this way. I’ve been doing this for many years now. Even with the Taiwan administration since late 2014, it’s two and a half years to this moment.

We already did quite a bit of cases. The case with Uber, the case with closely-held company law, and so on, are all organized in this cross-sectoral, multi-stakeholder kind of way of discussion. People are now generally aware of this motive. It’s not yet unknown, but at least I don’t have to explain myself over and over.

We did already make some new internal policies this way using this new kind of discussion format.

That might tie it a little bit to another question I was interested in asking you, which was, I asked you a bit about the negative role of automation and the idea of bots and propaganda, using tech for bad in that sense. What potential positive roles do you see for the use of automation in society in the future?

As with all automation, is to save people from doing work that are repetitive and that one does not learn anything from. The promise of automation is to make humans do human things instead of emulating machines. That’s the main promise.

Do you see any specific use for bots or messaging within the public that might be promising or bring potential benefits that maybe haven’t been used so far politically?

If you limit to chat bots on social media, for example, there is a civil society idea of a Line bot that tells fake news or rumors. Whenever you share a rumor article with that bot, it can respond and let you know whether it’s counterfeit or not.

Even if it’s counterfeit, it doesn’t mean that it’s censored or anything. It just shows a disputed content and a clarification from the other side. It’s possible for two piece of news to be each other’s clarification, but at least it shows you both sides of the discussion.

That’s pretty useful for all the rumors around food, around public health, around all sorts of weird things going on. It serves as a last-mile delivery to your Line account. That’s one of their concrete cases.

I’ve been really excited reading up on Zhende Jiade. I actually just added it on my Line and have been messing with it a little bit. I’ll probably be talking to Johnson soon. That’s great.

Another thing I wanted to ask you is to describe g0v a little bit. I’ve read a lot about it. What would you say either is g0v’s central mission, or the role of civic hacking in society to you?

Well g0v is a space, so there is no mission. Everybody brings their own mission. There is a vision, though. The vision is essentially the domain name. [laughs] The domain name means a promise, a commitment of something that says if you see a government service, and you’re not satisfied with it, you can always do something better, and employ the same website name, except the O becomes a zero.

It’s a private typing service for the public service. The idea is that you can fork, meaning that you take any government website existing service and make it better, and publish it under the g0v domain.

If g0v people release their ideas, their source code, and their creations under Creative Commons, then on the next procurement cycle, the government can tell it’s vendor is to incorporate these additions in. Then it becomes the official governmental website.

It’s the idea of forking, merging, and sending patches to the public sector. That’s the main vision. But how exactly to effect that concrete mission, it differs from project to project.

I think it’s a really creative, new-age, and progressive idea. I’ve been really excited from everything I’ve read about it.

Another thing I wanted to ask you about is your mission with transparency within the government. I saw that Simon Cheng brought Taiwan up to number one on the open data index, what you see as good about that, what the role of transparency is within government, and what their mission is.

The main idea of transparency now is that we want data to have its social value. Taiwan passed our FOIA law around 2005, which means it’s a relatively modern FOIA law. It doesn’t say much about whether the information must be released in a way that’s mission-readable or licensed under a free culture, open-source license.

That’s the additions that Simon Cheng brought to the table by essentially introducing regulations. Not revising the FOIA law, but interpreting FOIA law as saying it must be published in a way that’s not just human-readable, but mission-readable.

Not just readable, but also remixable, and there’s a threshold of any ICT system that builds below this amount of budget, like NT$50 million in the past three years must open data by default, except for of course privacy, national security, trial secret.

It’s a very umbrella mandate that basically means everything is open data, which is why Taiwan raised on index this quickly, because it’s built on the existing FOIA law implementation pipeline. We just changed the license. That’s Simon’s contribution.

My contribution so far is to make sure that the data itself is tied to a context, to a narrative, to its role in policy-making, and making sure that early-stage discussions between stakeholders are produced on this data. There’s no data, no evidence, then the ministry steering the hackathon are encouraged to produce the relevant data pertinent to the discussion.

One of the last things I’d like to ask you is you’re identifying as an anarchist while working in government, and trying to change it a bit. Could you talk a little bit about what seems to be that paradox on the surface?

The thing with anarchism is that I don’t take direct commands, nor do I issue commands. That’s the tenet of anarchists.

Of course, I can choose the much more acceptable way and say I’m a practitioner of direct democracy, but I think direct democracy differs from anarchism a little bit. Direct democracy is, of course, one part of it, but if you read the original anarchist text, it’s much more.

There’s a communal co-ownership of facts, of production tools, and the diversity an inclusion of all kinds of different people. This willingness to work with non-specific people is, I think, the key to anarchism. If you just say direct democracy, it looks like just referendum everything, which leaves out to inform the collaborative discussion part before, and guaranteeing the inclusive part afterwards.

I would still prefer the term anarchism. As for the apparent paradox, never issuing commands in my current role. Plenty of people want me to issue commands. I have resisted all the demands so far. It’s a pretty unique position in that I am a minister but I don’t have a ministry. Nobody really reports to me.

Of the 15 people who volunteer to work with me, most of their salaries are being paid by other ministries. They just want to work with me. They’re volunteers, essentially. They can leave any time.

It is this ad hoc group with standard consensus making mechanisms, business or economy, combines stand up meetings, scrum board, and make sure that everybody lists their own objectives in a classic OKR management fashion.

I don’t have any objectives of my own. My objective is to facilitate the team’s agenda. That also applies to the 40 or so people in the Participation Officers network. That’s pretty much it. There’s no paradox because I refuse to issue commands.

[laughs] It’s much more a team than a hierarchy, then.

Exactly. I call myself "public servant of public servants." 公僕的公僕.

I think it’s super logical and very revolutionary. The last thing I wanted to ask you is whether there are other people you think you could recommend about talking to the subject, automation, digital sphere in politics, the things we’ve discussed.

If not, that’s fine. I’m just curious if anyone would come to mind.

I’m not that active in the coding scene anymore or things like these. I think Johnson Liang or Ronny Wang, that’s the NewsHelper person, will have much better recommendations of their co-practitioners in the same field.

I’m definitely interested in the social side as well. Those are good places to start.

Once again, I really appreciate the work you’re doing in promoting transparency at high governmental levels, and also making time to talk today. Thanks for your good, progressive work, and parting out a chunk of time.

OK, looking forward to the recording. I’ll try to get a transcript done for you if...

Definitely. I’ll look that over, and then if there’s any sensitive info, I’ll let you know. But I think it should be fine. I’ll send you the recording immediately.

Thanks again. It was a pleasure to meet you.

Cheers, bye.

Bye.