The first thing I wanted to ask…Joel and I were corresponding about, of course, Taiwan being on the front line of democracy, as many observers have commented. I’m wondering though. Taiwan is less recognized for the innovative products that it’s put out to protect democracy.

Front line in advancing democracy.

The rhetoric of basically Taiwan’s democratic processes under threat and the defense of it at the frontlines, but it’s under-played, at least in the United States, how much Taiwan’s developed tools to protect its democracy.

You could also say that in Hong Kong authoritarianism is under threat.

Yeah. [laughs]

It’s the same rhetoric, however you want to say it.

Part of the research though that I want to highlight is how innovative Taiwan has been. Potentially, the recent election shows how successful it’s been in protecting democracy. A lot of the tools are not implemented in the United States that Taiwan has implemented. My first question is “Just what program are you most proud of as digital minister?”

The Presidential Hackathon. I can send you the transcripts that describe this. In a nutshell, it enables social sector innovators to partner with public and private to form data collaboratives on collaborative data governance around the likes of air quality, water quality, remote telemedicine, and whatever.

The best five ideas, every year, are co-created across three sectors. The five team receive a trophy from the president, which is a micro-projector that, when turned on, summons the president. The president shows her promise to implement whatever they have done in the past 3 months within the next 12 months into national policy.

It’s the binding power as a trophy to the Presidential Hackathon. We also have the international track. The winners last year where Malaysia and Honduras.

That’s awesome. I saw a YouTube about it. That’s great. Another question. Whether it’s an actor or a specific cyber threat, what do you think is the biggest concern for Taiwan’s democracy? What do you think is the biggest threat?

The biggest threat, of course, are in the digital domain. That is the potential for the digital reverberation of anger which is a kind of powerless sentiment into outrage, which is a powerful but destructive sentiment.

The sowers of discord use this kind of echo-chamber effect of social media. For every social controversial issue, there are, of course, five statement that could be easily identified that can divide a society neatly in half to radicalize the conversation.

Actually, there’s a lot more consensus. People understand that these are the basic things that we can not only live with, but really want something to happen. This is a real conversation in Bowling Green, USA about how to change STEM into STEAM. That is what the Democrat and Republicans all agree on using a online wiki survey tool called pol.is.

I think democracy, the main risk is if we over-focus on these issues and lose our sight on these issues, which could be pushed forward in a cross-partisan manner. If people lose sight on these and think people as essentially not humans, then it propagates the idea of what we call exclusionary populism. That is the main threat of democracy.

Do you see the media as basically cherry-picking these more divisive comments because they’re arguably newsworthy, but they don’t actually represent the rest of the society?

I don’t think so. I think the media, when they understand that they can co-create newsworthy items with the social sector on these topics, they work very well. In Taiwan, during the presidential election, we see thousands of volunteers working with media on various different position, initially, collaborative fact-checking in real time on the three presidential candidates, on every words they say.

Obviously, they can collaborate on these. These are newsworthy as well. The main thing was that the media need to undergo a digital transformation. Instead of seeing themselves as broadcasters, which is essentially speaking to million of people, they need to be co-creators that listens to millions of people, or more importantly, having millions of people listen to one another through the media lens.

Part of my research is looking at a lot of PRC actions, trying to better understand the United Front and their campaign. The Western media has focused on, in the tech sector, coordinated, inauthentic behavior.

Like fake accounts.

Exactly, like fake accounts. In the traditional media, like “Hongmei,” or “Red Media.” I’m wondering, as the digital minister, do you see the digital domain presenting other threats by PRC action, or are those really the two main areas that confrontation’s happening?

The point is that the two venues you describe are covert venues, essentially. We do also see overt venues, which are often not receive as much press. For example, there was a trending rumor on the Internet, that fact-check number by 204 for the TFCC, that says, “The rioters in Hong Kong are paying people $20 million to murder police,” which is quite an accusation.

It’s not only phrased in a content-farm-like picture card. There’s even so-called posters of so-called recruitment of so-called suiciders written in mock Cantonese, although with some titles, introducing people to telegram channels in Hanping for some reason. I don’t know why Cantonese protestor would use Hanping, so a little bit amateur. [laughs]

Wow.

Then the TFCC did attribution work. It turns out that the originator, the very first post having this counterfactual information, is actually from Zhongyang Zhengfawei’s Weibo account, which is the Central Political and Law unit, the Chang’an Sword, Chang’an Jian, which is overt.

It is not covert. It is essentially the state propaganda apparatus saying that these young people, 13 years old, pays each other 30k to buy iPhones in exchange for rioting. Actually, the photo is Reuters, [laughs] and it says no such thing. It’s a misappropriation. I’m sure that they will be very committed to structurally respect copyright more from now on.

This will not happen in the future.

Exactly, since the US-China trade pact, in phase one, it’s going to fix IP.

Yeah, that’s part of phase one, so maybe they’ll refrain from reusing Reuter’s pictures. [laughs] In any case, that is the point I’m making. Many of the originators do not take inauthentic account to mask their behavior. It is more like, in a United Front parlance, cognitive space-shaping. They use their overt accounts to do so.

The posting, was that from their website?

From Weibo, from Zhongyang Zhengfawei, Chang’an Jian.

Wow. Fascinating. Something like that, with the special relationship between Taiwan and Hong Kong, is that something that you see as directed at both populations? Or, have they also done similar work directly towards Taiwan and not the Hong Kong protest?

The thing is that they really have a cognitive-space incentive to paint the Hong Kong situation as a “domestic riot,” in which case the Taiwanese population will not identify with them as much. As everybody knows, the Hong Kong factor is the single factor in the Taiwan presidential election this time.

The cognitive-space shaping is maybe less toward the Hong Kong population, which can very easily see through this [laughs] so-called poster anyway, but less so, Taiwanese people apparently.

Another thing, which I don’t know if this was the think tank that you mentioned before, I was talking with a colleague at IRI. They were in Taiwan to learn more about Taiwan’s election as well. One thing we talked about was the public disinformation campaign in Taiwan in 2020 relative to 2018.

I wanted to tell you what we were thinking and see if you disagree. Basically, it seemed to be more muted in 2020 than 2018. Would you agree with that?

Definitely. I think the main factor is, first, people are aware that this is something going on. The fact-checking ecosystem is much more mature. Equally important, there’s less of a attack surface. Each referendum topic was a divisive point. In the picture I just show you, the 10 referenda topic are 10 points that divides the society.

This time, because we moved the representative elections and the deliberative direct democracy into alternating years, there’s less of a attack surface this time around.

Is it possible that there was other coordinated behavior that we haven’t picked up on yet? Do you think they were using maybe a similar playbook to 2018, but because they had less cognitive space to shape the discourse, just didn’t do it as well? Is there another venue, potentially, that is different than 2018 that potentially we haven’t fully seen the effects of?

That’s a DoubleThink Lab question. I don’t do that research. My point in working with the ministry is to make sure that we offer real-time, fast, funny clarifications to the fact-checkers. That’s the extent of the administration. The end you’re asking is basically for the academic and social sector to do the analysis. At least in our office, we don’t do that kind of quantitative analysis.

That’s helpful. Another question that I was talking with AIT folks about, has your office come across deep fakes? Did you see that used in this recent election cycle?

Not in this election cycle, but Facebook published a lot of deep-faked account photos in a mass campaign to create fake accounts on Facebook. It’s public, so if you look for “deep fake avatar photo Facebook,” you will see.

Interesting.

It really take a lot of expertise to look into those avatar photo and notice, for example, if someone’s wearing a glass, the glass reflects the light differently than in a real physical world would do. This is those small details. Otherwise, it looks just like a real person.

The inauthentic behavior can easily create tens of thousands of accounts that look all different. The narratives, the spelling habits, and whatever, could all be synthesized. Just this year, voice is now in the same domain as well. Previously, it was just pictures and mute movies. This year, voice is possible to do transfer learning on that as well.

Facebook have detected a lot of that. We don’t see those fake accounts identified by Facebook interfering with the Taiwan election, but potentially they could.

Exactly. That’s interesting. I’m also wondering, because the digital domain is relatively borderless, whether your office has come across PRC disinformation in other jurisdictions outside of Taiwan and Hong Kong. If so, what’s the next step? Do you coordinate with other jurisdictions?

We work with civil society in other jurisdictions.

For example, in Chulalongkorn University in Thailand, we work with their Communications Arts Department to bring the Cofacts people and the participation officers from the administration, as well as the private sector actors like the Open Data Alliance, all together to a similarly cross-sector conversation in Chulalongkorn University on social innovation and civic participation.

Disinformation is just one of the application of this ecosystem. As soon as we finished the workshop, we see Cofacts – if you take away the S so it becomes singular – cofact.org, you see the Thai version.

Nice, I didn’t know that.

We work with the Code for All system, initiated in the US. For example, their Japanese counterpart is Code for Japan, and I participated personally in the civic tech communities held for workshops on Tokyo and other cities by Code for Japan.

The Korean community also joined. I didn’t participate at that time, but the Taiwan g0v, the Korean community, and Japanese community held together the so-called Facing the Ocean Hackathon in a similar multistakeholder structure in Okinawa, which is equally close for all three [laughs] stakeholders.

The FTO, Facing the Ocean, is like a civil society-initiated GCTF. It’s very interesting to see it co-evolve with the needs of those jurisdictions to work together on disinformation issues.

That’s awesome. My next question was about…

Cofacts.

…Cofacts, yeah. How effective do you think both Cofacts and Taiwan Fact-Check Center have been at countering disinformation? Based on their successes, where do you see remaining gaps that maybe your office or you’ve been trying to engage?

In the mLearn website, the media competency, education framework, mlearn.moe,gov.tw, as you can see, the main focus items based on what they have done is first to translate it into K to 12, embed it into the curriculum so that the young people…

…can get media competency before they are of legal age to vote, obviously. Also, they focus on life-long learning. You see that the prominent journalists across various generations, they all contribute because they are trusted figures by various generations.

I’m talking about people like ???, or for even older people, maybe they trust ???, very noteworthy news anchors. The cross-generational translation of the media literacy messages to empower people on different generations.

FB, as I understand, is working with the Hondao, Elderly Care Foundation, to work this into elderly care home exercises and things like that. This life-long education, plus K to 12 education, empowered by research of higher education, journalism profession is the logical next step. They’re really doing a great job on this.

Has the program been well-received by especially the elderly and other adults? I’m just taking the US perspective. I don’t know if it’s so many differences in our society, but I would imagine [laughs] my grandparents’ friends being skeptical of being told what to see as fake and what not.

Generally, it’s been well-received and adapted by the elderly population?

Yeah. Generally, because the elderly also don’t want to always be the person in the LINE group in their family to be corrected, they also want to occasionally correct their grandchildren.

[laughs]

That elevates their social status within the family, so there’s incentive.

That’s funny. More recent development…

The AIA?

Yes, AIA. Of course, it’s part of the administration’s policies. I’m wondering though about the concerns mostly I’ve seen from Taiwan academia about free speech and potential overreach. Are those viable and serious concerns, or is it more overblown…

First of all, the AIA is not a Calvinist build. It is something strictly from the parliament. The opposition’s parties have also formed their own versions but there’s not much difference which is why they very quickly passed, did the act. The administration organ to implement the AIA is the MAC, obviously, mac.job.tw.

I’m just quoting from their press release saying that they’re working with the MOI, MoFA, the Minister of Justice, the CEC and so on to work into a taskforce and what they are doing first is that to make sure that people can very clearly see what fits in and what doesn’t fit in the scope of the infiltration, so-called infiltration. Their main message is “Don’t panic.”

They also understand that this bill is a first version, version one. They are also open to the possibility of version two or other complementary laws and regulations from other participating ministries. They will do so in a transparent fashion. They will make sure people see actual cases and how to tell what are infiltration in, what are not.

Finally, they will work with the head COA to setup a service window also for telling these people currently working or living within the patency territory, so that they can get the most frequently ask questions clarified quickly. That’s what the administration has been doing.

Do you think it’s an effective step forward, rather than something from the private sector? The US rhetoric is whether Silicon Valley can help solve some of these problems on their own. The Zuckerberg testimony in our Congress saying that Facebook was taking practical measures. There’s a pretty strong debate right now about how much we can rely on the private sector to do this and how much relies on government action.

Of course, there’s a heavy dose of skepticism whether the private sector will have the incentives to really protect US democracy rather than make profit. This has been going on right now with our political advertisements. Whether they’re allowed on social media, right?

Is this step–the Act–kind of a step that says the government, the legislature, has to take action and they can’t rely on the tech sector, or not necessarily?

We’re a bit ahead though. While you’re still debating the honest advertisement, [laughs] I think. We’ve already passed a law two years ago, to publish campaign finance and expense, as well as data, for independence analysis by investigative journalism so we have a pretty strong norm of complete transparency when it comes to election.

What we have done is to point at these norms established by the social sector and sent to Facebook, Google and friends, “Since you have signed on the self-regulation package for recovering this information, you have two choices.

“You either conform to the social norm around campaign donation and treat all your advertisements, social or political dream election session as campaign donation published in real time, in a structural data, that contains as many fields as least as the contributor fields because these are come in expenses or you refrain from running them. Your choice.”

It’s not even a law. It’s just a social sector norm. Instead of facing social sanction, they ought to choose either – like Facebook – to publish the Ads Library for the social sector to analyze in real time. Or like Google who referred from running political out skill in an election.

For us, this is a done deal. We have a norm. We don’t need a law. This is a self-regulation for the private sector. It’s based on the inclusive for the social sanction. I think the main difference with the US is that I currently don’t think the voting population have a very strong social sanction against.

Maybe they do have social sanction against people. We use the fake in-campaign. That’s a norm but I don’t know if that can mean donation.

Our Supreme Court said that’s free speech [laughs] and it doesn’t even need to be exposed as to who the donors are. That’s a good point. Did Taiwan force Facebook’s hand on the ad library? Because that is something that was instituted in the United States recently. I’m wondering, if the legislation in Taiwan is two years old, potentially, Taiwan was the impetus to do that.

It is to the FB’s interest if they can show that in the liberal democracy this helps the democratic process, so that when more authoritarian jurisdictions try to compel if we are to do otherwise and we can point to Taiwan as a positive counter example, and status and your interests.

Nobody need to force anything from FB because it’s obviously if there’s no positive example at all of add to library working, they would have a much harder time resisting demands from other authoritarian jurisdictions.

Related question: On the 2019 laws from the spring about “fake news,” both the legislations and the regulations, I saw some pushback at least from Google. I don’t know about other tech companies on whether they would censor or remove information that they saw as fake. How does your ministry interact with that legislation? What do you think about that debate?

The legislation is the responsibility of Minister Lo. I’m mostly the person who talk with those multi-national corporations. Anything that is norm, I help communicate. Anything that is law, that is Minister Lo’s work. That’s how we work together.

We had a very good conversation to the social media companies as to definition of disinformation at the end. It’s praised by the journalism community that we don’t use the term “fake news” at all. Because news, ?? and journalism, ???? shares the same word There’s literally no way in Mandarin to say fake news without offending journalists.

Both my parents are journalists. Out of filial piety, I can’t say F word. [laughs]

Disinformation is defined as intentional untruth to harm the public, not the minister’s image. If you set this file, all three, this clearly outside expression battery. The end result is something that social media companies can live with because it’s for the court to decide. Secondly, not the administration.

It is something that the journalism community really like because it plays more power to the social and academic, journalistic centers, not the public or private sectors. That’s my appraisal of it. It is not my work, it’s Minister Lo’s work.

That’s very helpful. I saw one of the debates…I honestly don’t know if it was resolved. When the regulation started coming out, one of the spokesperson from Google said…They actually cited their norms and their policies, and said that Google will not remove information unless it was sexually explicit, hate speech, or inciting violence.

The definition that they cited for “intentional harmful untruth” would potentially fall in between those three buckets.

Which is why they signed on this self-regulatory norm package. That is to say they agree to conform to their definition of intentional harmful untruth. They listed the actions they’re willing to take to do it.

Take down is not what we’re after, because take down is not useful in generally speaking variation is so easily propagated. Take that individually, we will overwhelm the court. It’s just the fact of the life, it’s far better for the likely virility to be bow down.

If Facebook recommended it, it’s like spam. We’re not taking down spam. We’re moving it to the spam folder. That’s one thing. Or, doubts. Whenever they see the information, they also see a short disclaimer that says NCC have fact checked this and this and it’s false actually. Click here to learn more.

That’s what we called notice and public notice. It’s far more preferable than take down. Take down is a judicial matter. Some public notice actually recruits more volunteers into the fact checking ecosystem.

The notice, or public notice, I think is accepted as a common norm between the various stakeholders and the points you see raise from Google at that time was to resists a nasty G like framework where they will be able to assume all the cost in Germany.

They really don’t want to do it in every other jurisdiction because it’s not very shareholder-friendly [laughs] approach. We don’t end up taking the nasty G approach.

Has Google and the government resolved their differences?

Definitely.

The line I saw was, “We’re not going to remove such and such…or censor information, unless there’s a court order.”

Disclaimers.

That’s disclaimers, OK, and that was the…?

That’s public notice.

OK.

The country disinformation norm package, I think their official document is only in Mandarin but there’s an unofficial English translation which Joel will send you.

Thank you. That’s very helpful. This is my last big question. As the United States enters its own election cycle, what do you think are the most valuable lessons learned from Taiwan’s 2020 election cycle for which the United States needs to take notice?

Don’t panic.

Don’t panic.

I think it’s too late for that.

Really, when people is in a panic, expect more panic. None of these social sector mechanism, which is very intricate, have room to grow. The more humorous or fun meme there is, the more room for real discussion around things.

You see this in the Better Reykjavík Icelandic conversations. You see it in many Nordic countries, where they were very successfully using humor to dispel the issues that could potentially divide the society. The initial formulations of Pirate Party is actually around very similar ideas.

We see worldwide that if the politicians can make joke about themselves…

[laughs]

…then it’s fine. It actually brings liberal democracy further. If they do the opposite, then people enter into a state of panic. That becomes much less possible for this kind of social sector-led formulations to grow.

With the work you’ve seen, again, across borders on Cofacts, for example, or Code for All, are you confident when you look at US civil society and fact-checking relative to Taiwan, or do you see the United States has a long way to go before we have the same protections Taiwan does?

My office is essentially in charge of working with the civil society on civic participation and social innovation. There was a office of that name in the US.

I, of course, think that there’s plenty of civil society energy in making that happen. The playbooks that came out of that office and now with GovLab, Beth Noveck and friends, which I’m a international advisor of, still encourage our work here. The US is a very large place.

The main difference when I talk to people in Toronto, where we also had a workshop, and in Ottawa is that you need to think about Taiwan more like Toronto and less like Canada. From Taipei to Kaohsiung is just 90 minutes by high-speed rail. 98 percent of people is on high-speed Internet. It’s €15 euros per month unlimited data connection.

There’s no digital divide in a physical layer. It’s mostly in the cultural layer. People feel that we are part of the democratic process even though we come from very different culture. We’re a transcultural citizens republic. All this shape a social norm that more resembles a municipal government in the federal system in Canada or in the system in the US.

Most of our learning is applicable to a similar slice like New York City. It would be applicable to New York City but less applicable when you’re talking about a federal system where people not necessarily feel they’re in the same polity. Some states feel that the state is a polity and the federal level is another polity and things like that.

That was very helpful. It was great. Awesome. Is there anything else that you think we haven’t hit on?

No, I think it’s really great questions.

Great. Thank you so much for your time.

Thank you.

It’s really an honor to meet you.

Thank you.

I really enjoyed your TED Talk, by the way.

I watched it on YouTube.

Thank you.

Thank you so much.

Thank you. Take care.

Take care.