2020-12-15 Meeting with Microsoft SLT

Greetings. Hello. Greetings. Is the sound, and voice — which is the same thing I guess — and video getting through? OK, excellent. That’s awesome.

Yeah, really glad to. We are definitely the future — just a few hours in the future. Feel free to make that trip to the future soon. [laughs]

I think I’ll leave more time to the Q&A than the initial pitch because most of the materials that I will cover is already covered in the written summaries. I do want to try the screen-share thing, just to make sure that…

Sure, of course. Please do. Yeah.

Yeah, I’m wearing an earphone. You can see that it only records when my side has the voice. When Glen was saying anything, there’s no voice here.

Really, I just want to share the cute dog, that’s the only thing I want to share. Do you see a cute dog? If you do, your Apache installation… [laughs]

Let’s get started. This is Zongchai, a Shiba Inu, literally a spokesdog of the Central Epidemic Command Center, the CECC, and the key figure in getting not only the pandemic countered with no lockdown, but also the infodemic countered with no takedown.

When the Zongchai reminds you to wear a mask, the Zongchai uses a very particular way of saying this. This is in Mandarin, but I hope that you can see the dog is putting literally foot in the mouth. This is saying in no unambiguous words that, “Wear a mask to protect yourself from your own unwashed hands.”

This is very powerful, because it appeals to rational self-interest. It works even when there’s no vulnerable people around. It works even if there’s no elderly people around. Had we said anything else about respecting others, this would not have a R value above one, I mean this message.

Because this sets self-interest in a rational way, in a very cute way, it appeals to all political inclinations, to all age groups, and to all nationalities when we translate this. People actually wore a mask and washed their hands, because this is actually in their best interest doing so.

There’s a system doing this, remember to cover your mouth and nose when sneezing, when outdoor, keep two Shiba Inus away from one another, otherwise wear a mask. When indoors, keep three Shiba Inus away from one another, otherwise wear a mask, and so on.

There’s name for this technique. It’s called humor over rumor, and indeed when we work on the counter disinformation work, we are constrained by a lot of people still remembering the martial law, me included. We don’t want to go back to censorship, to administrative takedown, because it’s really bad.

Just like when we counter the pandemic, we don’t want to go to lockdown, because we understood in 2003, when SARS 1.0 was released in Taiwan, there really is [laughs] no good thing that will result from a lockdown for economy and also for the mental health of the citizens.

This time around, we have to do this with no lockdown and takedown. When we detect there’s trending misinformation, we discovered that instead of taking things down, which enrages people – and this misinformation treating outrage anyway, and that will further increase our value, we might as well just use humor.

We started prototyping this humor-over-rumor thing before the pandemic. A few months before the pandemic, our Premier, Su Tseng-chang, the Head of the Cabinet, has always been sharing that there’s a lot of meddling to undermine the democratic process.

It’s always the CIA, you see. “The CIA made two invisible ink for ballots. No matter how you vote, the magic ink always shows Dr. Tsai wins” — something that of course doesn’t happen in advanced democracies like the US. [laughs] Anyway, we do have the disinformation going around trying to undermine the January election.

We countered them not with takedowns but rather by inviting YouTubers to watch the counting process. Our counting process is transparent. Each person takes out this one ballot, it’s paper ballot only.

Then the audience watching can actually freely just start tallying themselves using the recorders. Each side actually has their own counting app. When the YouTubers and the counting apps agree and there’s video footage for pretty much every counting station, there really is no room for the rumor to grow because everybody can see in a very fun way.

We have a turnout of around three-quarters, 74.9 percent. With that kind of turnout, we can actually work through participatory accountability during the counting process, to publish the early signs that could be disinformation or misinformation, and work with the post-partisan YouTubers.

Regardless of parties, all the four major parties here agree that deepening democracy, meaningful international collaboration is important.

In addition to those civic technologists and people in the K-12 schools learning about media competence and cutting their teeth on fact-checking the presidential candidates and the counting process…

We also have government ministries that detects those end-to-end encrypted channels for trending myths not yet disinformation and wrote out funny vaccines of the mind, that is to say those humor over rumor memes, to counter it.

We don’t, of course, have a visibility to end-to-end encrypted conversation, like on the LINE or WhatsApp. What gives? We learned a page from the counter spam. Anyone can long-press a message saying this is abusing my trust to my friend because my friend is evidently not just simply forwarding something that trends on outrage but not fact-checking it.

Anyone can long-press something and send it in an anonymous fashion for the professional fact-checkers to work on. There’s a trending dashboard that’s open data actually. Anyone can see what actually is a trending rumor.

We focus our fact-checking energy, both in the social sector and also the public sector, only working on the things that look like it has an R value above one and ignore everything else because they will not go anywhere anyway.

We have this triple-two rule that says for all the competent authorities or the ministries, if you see a trending rumor still at the misinformation stage, then on average they now roll out a meme within 16 minutes. The deadline is two hours. Within two hours, they have to roll out two different pictures, each 200 characters or less.

This is one of the early example. You’re seeing our premier, head of cabinet. He says, “There’s a popular rumor that says if you repeatedly perm your hair within a week, you’ll get a $1 million fine. That’s not true.” A younger version of himself saying, “I might be bald now, but I will not punish people with hair.”

The fine print says, “We have introduced a labeling requirement that takes effect on 2021 for hair ingredients.” With a hair blower, a premier, as he looks now, says, and I quote, “However, perming your hair multiple times within a week will not damage your bank account, but it will damage your hair. Just look at me now for what will happen to you.”

Again, this is hilarious. This is absolutely rival. Sorry. Viral and also rival. When you Google about perm hair fine or whatever in Taiwan, you’re bound to see these hilarious pictures instead of the original conspiracy theory.

In this way, we make sure that the people who associate the long-term memory after a night of sleep of those keywords into something hilarious and also prosocial, co-creative, and things like that.

That’s our counter-disinformation playbook, which by and large play a large role in also countering the pandemic by basically having the daily press conference with the quint, the medical officers. They don’t look like professional comedian, but there are teams of professional comedians and participation officers who engages the trending hashtag.

One example and I will stop. Back in April, in the daily press conference, you see all the quint, all the five person, wearing pink medical mask regardless of their gender. The reason is that one night before, on 12th of April, there was a young boy that called the toll-free number, 1922, complaining that we’re rationing all mask, of course.

It just so turns out that all he gets is pink medical mask. He has lots of pink medical mask, but he doesn’t want to wear it to school for he is a boy. All his boy friends have blue medical masks. He doesn’t want to go to school.

The call center people in 1922 doesn’t know how to answer that. Although they have immediate answer for 95 percent, of course they don’t have an answer for that. It gets escalated immediately to the CECC.

Then the participation officer suggested two things. First, all the medical officers need to wear pink for the next week. Next, the minister, Chen Shih-chung, the Health Minister, he even said the Pink Panther was his childhood hero or something. The boy became the most hip boy in the class, for only he has the color that heroes wear and the hero wear, I guess.

This rapid response system actually shows that all the social innovators, they could be a professor, Lai Chane-yu, who just invented a way to use traditional rice cookers. Without adding water, you can cook the mask to destroy the virus, but it doesn’t destroy the mask.

Or any of those social innovator that make the mask map and so on can just as simple as calling 1922 and then get escalated to tomorrow 2:00 PM’s press conference, invited to explain their ideas.

Trusting citizens with real-time open data, which makes it open API, so that people queuing online can actually witness the person queuing before them swiping their national health card, which covers 99.99 percent of citizens and also residents, immigrant workers too.

Actually get to see on their phone that there’s children’s masks, which used to be 196 on their pharmacy, actually goes down to 186, meaning that they completed a batch of 10 medical mask per two weeks retrieval. This is participatory accountability. With more than 100 different tools, this is essentially a distributed ledger.

What we are seeing is that people trust each other much more when we do provide these kind of 30-second updates. If we update every week or something, like a Freedom of Information Act, then of course nobody trusts each other because there’s no way to participatory audit that. Even if we publish every day, you still have to trust the national health insurance.

With real-time auditing, if you see rather this number increase, then you will call 1922 right there. This shows something about data coalitions, which is, of course, built upon our existing data coalitions around climate, around water quality, and also around air quality. This is entirely social sector.

People identify gaps in those industrial parks, so they negotiated with those entirely social sector-built air quality sensor for the municipal government to start working with them on offering the labs within the industrial parks which is owned by the municipality to join their distributed ledger in measuring climate and also air quality.

We have the civil IoT system, ci.taiwan.gov.tw for that. I think I’ll just stop here — instead of introducing anything like quadratic voting — because Glen can do that in his copious spare time. I’ll stop here and take some questions.

There’s three pillars – fast, fair, and fun – and there’s different ways of scaling. You can scale up, scale out, or scale deeply.

I would say that the fairness part is the hardest to scale, because for the mask rationing to work not only do we have to have a single payer, universal healthcare system that makes visiting pharmacies and clinics cheaper than drive-through test, but also have to have broadband as human rights, as well as all sort of things. That’s the fairness part the participatory accountability part is built upon.

On the other hand, we have scaled that out, nevertheless, to South Korea, which is a country with more population and more geographical diversity.

The civic technologists there, as young as 14 years old, I videoconference with them, saw our mask rationing stuff. They managed to build the same system without having a single universal healthcare system running in the pharmacies by doing a lot of workarounds and hacks.

After a month or so they wrote out the mask rationing. The same map from Tainan, Finjon Kiang’s map, works in Korea as the first available map, even though Finjon doesn’t speak Korean. He speaks JavaScript.

It is possible. It’s just a lot of work. [laughs]

On the other hand, the fun part is very easy to adapt. We introduced that to Thailand, to the Chulalongkorn University.

Before long, the crowd sourcing, fact checking thing, cofacts.org, they removed the S, become singular, and registered cofact.org. If you go to cofact.org you’ll see a Thai version of the Taiwan social sector crowd source fact checking.

That’s very, very easy to copy. Even though they’re not initially fact checking politically related messages, because of different scale of freedom, we saw them building their solidarity with the unions and consumer protection organizations by focusing on health, drugs, pharmacies, and so on. Then we eventually saw some other mobilization, from which I will carefully move on.

I will also say that the fast part, the collective intelligence, may not be so essential to do on a United States level. I consulted with, for example, Beth Noveck in New Jersey while she was doing this “Ask a Scientist” portal, which is quite like our 1922 call center.

It could scale to New Jersey, not necessarily the entire federal government, but that’s not necessarily a bad thing. For a lot of things, like crowd sourced agenda setting, it relies on people having the same shared reality, instead of living in their virtual realities.

It’s easier to start small and start in a community, like the Polis, which has been tried out quite successfully in Bowling Green, Kentucky. If you start by saying the entire United States, then persons in Hawaii, persons in Alaska, and so on, may have different priorities and realities. I would suggest you start small, and that’s it.

Excellent question. Yes, definitely. We have tried some sort of bilateral scaling of this with pretty good success with the AIT, which is the de facto embassy of the US to Taiwan, in what we call the Digital Dialogues.

The Digital Dialogues, they are not toy problems. They are actual issues.

I will share the website. Hopefully you can see it. It’s called AIT@40 Digital Dialogues.

It talks about, for example, how to get Taiwan more uniquely seen in the global community, how to remove the trade barriers and have a prosperous network in economy, how to promote US-Taiwan security cooperation, and how to promote people-to-people ties. You can the Polis and Dialogue results later.

The point here is that this talks about the common values, despite different cultures and different possessions, with a set of questions. The binding power being the people of both sides, after they share their common values, we use these and only these as agenda for the diplomatic corps, the foreign service of both sides, and in the case of the economic one, also the chambers of commerce and so on, to respond point to point in their white paper about the commonly identified ideas.

We had a pretty good success, so we would later on work on the Cohack, which is the collaborative or coronavirus, your choice, hackathon. That has more countries participating.

There were some outlandish-sounding, to Japan and Taiwan, ideas. For example, people using assistive intelligence to prioritize people going to ICU – who has the more utility to get faster treatment. This is completely out of norm around here, in Taiwan and Japan, but we have our share of our counter-norm ideas, as well.

Also, about data. This shows the three different regions show three different data norms.

Nevertheless, Polis eventually allowed us to see the common ideas. For example, if you record your health information, keep it to your phone, and only generate one-time link for the contact tracers, that is billed, then, as a privacy enhancing tool because, unlike a traditional interview, which will divulge more private information about your friends and families, this works in what we call real contact system, rather than real name system.

This won the consensus of all the five countries that participated. For example, Gemini Explore here allows your top epidemiology expert to convince better your vice president. That also won acclaim.

In Taiwan we don’t quite need that because our top epidemiologist expert, when he want to talk to the vice president, just looks into the mirror, but it is very helpful to other jurisdiction.

To focus on the common norms, to use a, what we call, people-public-private partnership to build the norms first, and then scale that out with the business sector, that seems to be the winning recipe. That will allow cross-jurisdictional, and even transcultural dialogues.

I remember the martial law in Taiwan, which is a quite low beginning point when it comes to levels of democracy. The same kind of participatory accountability is already in the works in the social sector, and to a lot of success, including the earlier co-ops movement, and things like that.

I don’t think it requires a liberal or social democracy to work. It does require some sort of literacy in the reading and writing sense, numeracy in the sense of that if you look at the Polis graph of the principal component analysis of the main contention points, as well as the k-means clustering of the main ideas being grouped, then it need to make sense to the person looking at it.

It could be in an analog form, like open space technology and things like that, but it requires a certain sense of numeracy to see this is the actual UberX conversation, that personal liability insurance very important. If you agree, you move towards me. If you disagree, you move away from me.

There need to be some sort of numeracy that makes sense of this movement, which will then enable the most important picture, which is the Polis report. That every time we run a Polis conversation everybody see that the divisive statements are just five percent and mostly everybody agree with most of their neighbors on most of the things most of the time. This flips the antisocial social media around.

For this picture to make sense some literacy and numeracy is needed, but nothing more than that.

Definitely. In design thinking terms, to overuse the Double Diamond metaphor, our techniques is very, very good at expanding, literally, the degree of the discovery stage so that it will cover a very wide swathe of people.

Also, help on the define stage, because for the career public servants to do this they are limited by the bandwidth of the state. That is to say, their qualitative time to spend on those ideas.

Now it’s the crowd moderating itself, so it scales automatically. Even if you have a very wide diamond, it eventually get it defined semi-automatically in the assistive, collective intelligence fashion into common values.

That only gets you halfway toward the “how might we?” question. Everything after that, the development, the prototype, the delivery, and so on, are for professionals.

My pitch to the Legislative Yuan members, to the members of the legislature, is that it makes your life easier because you get to focus on the things that people actually care about.

This is like talking to the fact checkers and say, “Our system help you to focus on the rumor that actually have a R-value above one, otherwise it would waste all your time talking on those R=0.5 rumors, which doesn’t need fact checking to begin with.”

This is an early diamond system, but we do not presume to say that it replaces or in any way replace the expertise that’s required on the second diamond and beyond. With the open API everybody can unlock the final delivery. The delivery is, in a sense, just a prototype, but that’s another story.

I keep calling this assistive intelligence for a reason, because for an expert’s, a glass is of no threat to them. It just helps me see better unless this glass starts showing pop-up ads, in which case, it is a threat. [laughs] As it keeps on being assistive, meaning aligned, accountable, then of course, it’s no replacement and therefore no threat for the experts.

Yeah, we don’t have to think about it. We just face it, everything. Quite literally, we are on the front line. [laughs] In addition to this random “CIA invisible ink,” we also have not covert — really overt — attacks.

For example, right before our January election, around mid to late-November, one of the most trending rumor on Taiwanese social media is, “Hong Kong thugs are being paid 20 million to kill a police.”

This is obviously because whoever spread this understood that this is going to be the deciding factor for our presidential election. They want to distance the Taiwanese people from the Hong Kong people by inciting a sort of outrage that will then turn into revenge.

This is an actual photo that it used. It’s a true photo by Reuters. The original caption only said they are teenage protesters. Period. The rumor that makes the rounds is something like this 13-year old thug bought new iPhones, and so it’s a very different captioning.

The fact checkers, which are not organized or paid by the government, eventually traced this message to the 长安剑, the central political and law unit of the Beijing regime and is their Weibo. They very openly say it. All that trending disinformation came after they said something open about it interestingly only in Taiwanese social media and not in Hong Kong’s social media.

We didn’t take that down either. We just made sure there is a public notice. Whenever you share it on the social media platform that have signed on the counter disinformation accord, it says very clearly that according to a Taiwan FactCheck Center, this is Beijing-sponsored propaganda.

That is the way to counter this tactic that they use without becoming something like them. This is just one concrete example. There are many, of course, other examples.

However, just as in epidemiology, it tends to be that different variations of the SARS 2.0 are competitive against them themselves as well so that we are having only one strain [laughs] at the moment in our borders, and we need to focus our energy to fight that strain, not the earlier strains of the same variation.

Fortunately, the total cognitive bandwidth of people who are likely to believe conspiracy theories and share them is quite limited. Within a news cycle, there’s at most maybe two or three trending disinformations, and then we have to focus on these.

If there is a DDoS, that actually pushes out the mindshare of those less-than-viral disinformation pieces, in which case we don’t have to work on it as well. That is the basic idea.

Yes. Just to reduce its basic transmission rate.

Yeah. There’s an art part of this, but there’s also science. We in that mostly partnered with the platform economy players, such as LINE. LINE is a really good example because they also don’t want to be seen as the communication instant message platform that all the rumors began. It’s not a good branding. [laughs]

LINE is quite helpful. They built their own dashboard as you can see here, so you don’t have to rely only on the civil society to piece these things together.

The business sector says, “As of today, there’s this many reporter’s disinformation or misinformation. There’s this number of people joining. Thank you for volunteering. The trending ones are this, this, and this. These are the trending ones that are fact-checked. You can check it out,” and things like that.

Anyway, this is the basic idea. Now, not all of these information are false. You see, this says that “if you don’t wear a mask, in convenience stores starting December, then you may be fined if you keep refusing to wear a mask…” This is actually true. [laughs]

This also serves as a way for the people who work with those Shiba Inus to look at the viral form of their public service message and learn from it, showing that if we package like with RNA strands of this viral disinformation, then it actually reaches more people. Why don’t we do that?

There’s three sectors going on. LINE works with the social sector fact checkers, including Taiwan FactCheck Center, MyGoPen, who are all partners in the International Fact-Checking Network. There’s a larger social… It’s like Spamhaus, right.

Then the public sector’s role is just to making sure that first when a fact checker assesses information, we need to give it within 60 seconds. Also, when they check information from us, we know that it must mean that there is a trending rumor about this.

Our own comedians literally just walk back home because the person who is the Participation Officer in the Minister of Health and Welfare lives just, I think, one block or two from the Ministry building, and they live with that dog. [laughs] They just walk back and ask the dog posts some new model pictures and just take pictures of it and call it the Taiwan Model. [laughs]

I was like negative five years old, or something, in the age back then. [laughs] Necessarily, I’ll be talking about anecdotes.

In the late ‘70s, democracy was without form and void, and we needed it at the time. There’s absolutely no press freedom, so journalists sometime worked with international correspondents stationed in Hong Kong, to get the human rights violations in Taiwan out to the international like Amnesty International.

Without international attention, it would deteriorate very quickly. We rely on international watch and support. Of course, nowadays, this Hong Kong relationship is entirely reversed. That’s another story.

It just turns out that in Taiwan, the KMT government was following this playbook from Dr. Sun Yat-sen, who is a Henry George follower and believes in mechanism design.

There are parts of the constitution that is very idealistic, that talks about, for example, constitutionally, the states need to put a limit on business capital and encourage instead cooperativism, things like that.

There are languages in the Constitution that basically points out — even during the Martial Law — there are things that are considered sacred. For the thoughts of Sun Yat-sen is taught like a state religion — not unlike Maoism — in the martial law days.

The people who want to form a social sector simply followed exactly as prescribed in the Constitution to form the co-ops, to form the consumer right protection campaigns, to form the environmental understanding and awareness campaigns… In a sense, they are calling themselves social innovators. They are not necessarily “the civil society” for the same reason that PRC activists in Beijing don’t call themselves “the civil society.” Social innovation, that’s fine. Cooperativism, that’s fine. It’s fine as per the Constitution.

They started forming these grassroots charities, or organizations of social entrepreneurs that eventually come to essentially fulfill their role where the top-down government cannot see and gain a lot of legitimacy in terms of education, especially for the rural, and long-term healthcare, especially for the remote islands and mountains and so on.

Before long, when there is a natural disaster, like the September 21 earthquake around the turn of century, the social sectors’ numbers are widely much more trusted than the government’s numbers, partly because in 1999, which was the year around the earthquake, we’re still only three years into the first directly elected presidential term.

The government needed to catch up in terms of legitimacy compared to the social sector. Later on, when we occupied the Parliament in 2014, we made sure that the 20 NGOs that organized this Occupy, almost all of those NGOs were around for a long time, some even before the democratization. They serve as the bedrock of legitimacy.

We, the civic technologists, mostly connect those NGOs together and make it cool again for the teenagers to start talking about those 20 different things.

Of course, it’s described as a protest, but really, it isn’t just that. In the three weeks of peaceful Occupy, is a demonstration. In a demo sense, we demoed a way for the 20 NGO to hold an open, multi-stakeholder conversation about a trade deal with Beijing.

Half a million people on the street and many more online decided collectively that, for example, the economic argument for using PRC Beijing components in our 4G infrastructure doesn’t make economic sense because for each firmware upgrade for each generation of radio technology, we’ll have to do another risk assessment to see whether it has been de facto taken over by the Communist Party, and we have to keep doing this.

If we have to keep doing this, then it’s more expensive than developing our own or working with Nokia and Ericsson and so on. That’s just one of the many consensus formed during the Sunflower Movement Occupy. This derives directly from the origin of those NGOs during the martial law era, but in a sense, they are also compatible with Sun Yat-sen’s thoughts of cooperativism.

Yeah, about the resistance. Not from any of the four major parties… Partly because deepening democracy, meaningful international collaboration, these are just bread and butter of Taiwanese politics. Since 2014, anyone who opposed open government doesn’t get elected as mayors. It’s this simple.

We don’t get any resistance or opposition because, in a sense, we are the resistance. [laughs]

We do get some resistance from, of course, international commentators. For example, when we introduced the participatory self-surveillance, we do get some critiques.

I do receive emails from upset academics, quite literally every day during the height of pandemic — now just every week — asking me to explain, why the hell there is no opting out.

When returning to Taiwan, your phone needs to be placed into this digital quarantine, where the telecoms can just through the signal tower strands to get your rough whereabouts. And if you break out of that zone, then you get a SMS, and the local medical officer also gets a notice.

When they check your whereabout, if you’re not home, then you’re fined up to 33K US dollars, which is a very heavy fine for something that you cannot opt out. There are, of course, people asking, “How the hell is this constitutional?”

We simply pointed out that right after SARS 1.0 in 2004, our Constitutional Court told our legislature that the unannounced lockdown of the Hoping Hospital is unacceptable. Figure out something that’s constitutional and proportional doesn’t require GPS or Bluetooth or whatever other technology but will still work in a way with equity.

So this measure is deemed constitutional in Taiwan, but I’m not pretending that this sounds like liberty- and freedom-preserving for everybody else. On the other hand, it does make sure that when apply with equity, people feel it’s fair. It’s fast. If not particularly fun… We make sure that they have streaming video during the quarantine. That enabled the rest of the country to enjoy Pride Parade.

Yeah, definitely. I had a lot of conversation with people in Germany, in France, in Switzerland, and so on, and the point keeps coming up. There is basically we having the hindsight of — pardon the pun — 2020, looking back at 2003, looking at SARS 1.0, and saying that we don’t want to go back to lockdown because of this societal inoculation.

Much like in Taiwan, people around 40 years old and older don’t want to go back to the martial law. People around 30 years old and older don’t want to go back to lockdown. This is just that. In a sense, many people around world are being exposed for the first time a release of SARS, and they also need time to develop this kind of antibodies of the mind.

Of course, after the vaccine and assuming there is no SARS 3.0 or SARS 2.5 service pack, then of course, they will get some breathing room — quite literally — to work out the constitutional limits.

I have a lot of time. I have 40 minutes, but you may not. [laughs]

Do visit Taiwan. We have this social innovation partnership summit in April, and all of you are cordially invited.

Thank you.

Live long and prosper.