 Is this thing on? It is. Let's see if the projection is you. Here we go. Hello, everybody. Good morning. As a customary to my speeches and keynotes, this will be a crowdsourced speech. For those of you who have cell phones or any other devices that can connect to this great thing called the internet, please go to this website, which is slido.com. Once you're on slido.com, you can enter this event code without the pound sign, 01017, today's date. And once you're in, this is a anonymous chat room. The chat room basically enables anyone to ask me anything during the keynote and also like each other's questions. And I'm committed to answer every single one of the question. But if there's just too many, then I would just answer the ones that are of the higher score. So if you click like, the ones with the more like appears on the top, which I will highlight and answer. So while you're still maybe thinking about a question to ask, let me give you a very brief presentation about five minutes long, about our approach on this thing called misinformation, disinformation, and other cracks to our democratic governance system in the age of the social web. There's a famous line in a poet and a singer, Leona Cohen, that I very much like. It goes like this, quote, there is a crack in everything and that is how the light gets in, unquote. And here we're facing a crack in our democracy through disinformation and through misinformation and a different kind of way of undermining trust in our democratic institutions. But that also gives us a chance for the light to shine through and for global cooperation. Because misinformation, you see, is a phenomenon, but it gains a new playground in the age of social media where it is easier to share a piece of information than actually reading through it, where it is easier to actually make a meme go viral before anyone has the chance of commenting on it. So it is a very different age and very different atmosphere compared to paper-based media where you literally have to finish reading the pro and cons, the different sides that they report before actually sharing that piece of paper. So it is the screen-based reading really cause for a very different way of balancing information. And this targets, of course, not only the trust on democratic institutions or the governments, but rather it undermines the trust of everybody to everybody. The fake accounts, the various other ways that HASA actors have been able to use the social media and other online venues, undermines the trust that people appearing as someone else and people with deep fakes appearing as someone else in very different ways compared to our face-to-face conversation and face-to-face modes. And within all this, Taiwan upholds the freedom of speech as our core value because I still remember the days in the martial law we only gained true election and freedom of information and assembly and so on for one generation. So let's do remember how is it like to be without those freedoms and we are committed to keep those freedoms alive. And so we're kind of forced to innovate, to face this misinformation campaigns and this information threats with ways that does not destroy the freedom of information or freedom of speech because that is exactly what those actors want from us. So we basically, our approach is proactive instead of reactive, instead of waiting for misinformation to spread, we give a timely response whenever there's misinformation about anyone. Take a personal example, two years ago, just when people learned that I'm going to become the additional minister, there was a large parade on the street and there's a misinformation that spreads that says Audrey Tom is able to monitor everybody's whereabouts using GPS satellites and mind control. This, obviously it's not true, but it gives a symptom of the underlying distrust of people and location tracking and other legit suspicions about technologies. And so I took that chance to record a live video and also write numerously online and responding in real-time social media, basically proactively sharing what kind of digital governance am I envisioning? What do I actually mean by radical transparency? How people can actually find me in my office or in the social innovation lab every Wednesday and talk to me face to face and also publish our interactions online. So that gives me a perfect chance to outline the kind of radical transparency and radical trust that I wish to instill in my office. And so by responding proactively, I think this makes it less likely for misinformation to grow into this information. If on the other hand, based on those rumors, based on those speculations which are perfectly valid in your democratic society, the government, the institutions has remained silent, then after a period of silence, it creates the perfect opportunity for hostile actors who want to sell this court through intentional disinformation campaigns. And that creates the real problem because then it creates a fertile ground for those buyers of the mind to spread and to harm people's trust on each other. And of course, during election season, which is very close now, we also see people who even are willing to risk criminal offense to interfere in election and to intimate the public. Of course, we have legal code to enforce all those role of law when people commit criminal offense, but as people have already pointed out, how to automatically enforce these and how to expose those criminal offenses to the public, that is something that we can brainstorm and we can innovate on. So to very briefly recap, the perfectly valid and normal thing of rumors and sometimes misinformation, we work actively in a way that includes everybody in decision-making. For all the people who complain, for example, about our tax filing system last year, instead of just responding to those accusations, we invited everyone who complained a free pass to the co-creation workshop to make this year's tax filing report and our system. And I think this is the perfect example because anyone who participates actually understand how is it like to be in the kitchen of policymaking and rulemaking. And people are armed with an extra contextual information that they can act as vaccines of the mind so that whenever they hear something that is just partially true or mentally untrue about a tax filing system, they are able to spread the truth and to participate in co-creation. And this in turn, in constant accountable in transparent institutions, these are all sustainable development goals targets. They make wonderful icons. And for this information, and that is to say intentional spreading of this court, we work partner with the independent fact-checking mechanisms, both abroad and also domestic. We work as the introducer has mentioned, I work in the K212 curriculum and we're one of the first, if not the first place in Asia to introduce in our K212 basic curriculum, the ability to critically think and the capacity for the teachers to learn along with students navigating various information online. And so media literacy is one of the nine core characters or literacies in our new curriculum. And with that, I think a new generation that does not believe anything just because it's printed or said in some way in an authoritative way, but actually learns to navigate the various different layers of message-setting and can contribute substantially to the civil society originated fact-checking efforts. And finally, we need to provide psychological safety online. So people can read the short, that if there are actors who can be in the public or commit other criminal offenses, that the government will respond in a timely fashion to expose the actual intent and the degree. And if these are rumors or disinformation that we cannot explain because it occurs in an end-to-end encrypted channel or whatever, we trust our civil society partners to develop automated thoughts and other technology inventions that surfaces those dark rumors into the public where we can then tell whether it is just misinformation, people who are misinformed, which we then resolve just by proactively opening up our decision-making process. Or there are really hostile actors behind it that are working in a way that is concerted in which we develop vaccines of the mind, such as meteorology programs and fact-checking programs. And finally, if people are willing to commit criminal offense mostly to interfere in the election, then of course, timely judgment from the judicial process. And so finally, none of this, of course, can be done domestically. I am not even sure domestically, even make any sense at this point now because everything is widespread, everything is global, including the threats and the innovations to counter those threats. And so, using this very nice icon from the Sustainable Development Goals, we need to encourage effective partnerships. And so, that is my introduction. And let's see what you have to ask. So, feel free to, oh, that's great. We have 10 questions and have an error. Feel free to just add to it the new posted questions will appear in the bottom as the latest question. So, Jess, Macy, you would like to know, how does Taiwan know that fake news that this trace to China is also linked to the Chinese government? This is a great question. Not all IP addresses that traces back to any provincial or domestic origin is necessarily linked to the government, right? It can also be linked to the party or to the military. I mean, there is no clear indication that, of course, one of those three actors. But we can, of course, make what we know public and rely on independent and investigative journalists to complete the puzzle. So, instead of indicating any particular government branch or military or party, what we actually do is we share with the independent journalists what we have, our hunts, what facts, what evidence it's what data we have and the limits of which we also disclose. And we rely on international collaboration framework to piece together those puzzle pieces to give them much more complete picture than we can obtain our own. But we're totally committed to share our part of the evidence and the data that we have gathered. Many would like me to share my hobby, which is troll hugging. This is completely unexpected, but I would answer it anyway. So, yeah, my hobby is troll hugging. I shared this hobby about 10 years ago in a blog post. Trills, as you know, are people who just post misinformation and sometimes very toxic ones online in order to gain attention. Trills are not always state sponsored or even organized. Mostly, they were just people who crave for attention because they don't give each other sufficient hugs, I guess, in the real world. So, they really want attention from a monitor for some reason. And the trolls usually make maybe 100 words at hominid attacks, sometimes transphobic or sometimes, like, always personal toward me. And I've seen a lot of those trolling posts that's directly to me personally. And so, the troll hugging is a two-step movement. The first step is that if I see anywhere at all that makes me upset, that reaches me through the visual field that makes me feel upset, then I make some other feel in other sensory stimuli like make good music or to have a fine tea or something that smells good. And basically re-associate this visual word stimuli with something that is pleasant, that is coming from listening or from smelling or from eating. So, that creates a new associated in my brain so that the next time that I see those toxic words, I actually feel really good. And instead of being upset, which is exactly what the troll wants, right? And then I engage with the troll if and only if I can see maybe three words, maybe five words of the 100 word posts that are authentic. They review something of the troll themselves. They review something that's substantial that is of public interest. And they reply with a very calm mind very carefully to those three or five words. So, this creates two facts. First, that it tells the troll that they can only get attention that it craves if they contribute something to the public forum, the public discussion. And second, it tells all the bystanders that it is actually okay to review authentically part of one's experiences, maybe negative experiences because personal feelings, there's no right or wrong. It's okay to review those feelings and I do that as part of my responses. Sometimes I just record a short video clip. And so after that, the trolls learn gradually to review more of themselves because they understand now the attention that they're getting this way from the community and also from me is relational in nature. Previously, the troll only gets transactional attention. So they wake up and because the attention they got was transactional, they have to find somewhere else to troll and get a fridge batch of junk food because it's not really healthy to one's mind to have a series of transactional relationships where there's no interrelational reactions from anonymous people to anonymous people. So they've learned that there are some stable relationships that's possible online. And most importantly, I have this office hour in the social innovation lab. So actually after a few rounds of exchange, welcome the troll to my office hour every Wednesday from 10 a.m. to 10 p.m. in the social innovation lab which is like this. And sometimes they see self-driving tricycles drumming around. This is a very interesting place of co-creation and I physically give them a hug as well. And so this is troll hugging. And it would not be possible without the first a kind of cognitive behavior therapy on one's own mind. But after that movement, after that initial making room in one's mind, we can take a humorous and even lighthearted sound. One of the most difficult challenges in our world which is it's very easy to project whatever people want to project into a piece of words or a bit of image that's taken out of context. But after getting people in a habit of if they review something authentic that contribute to public discourse, they always get something authentic from the minister and it's welcome to face-to-face conversations to give a hug and things like that. I think people generally feel much better and can be more active contributors to democracy. Dr. Gavin Ellis from New Zealand would like to know what research is being undertaken in Taiwan to detect and thereby counter deep fakes and other AI generated forms of fake news. This is a hot machine learning topic. And so I'm sure that people will eventually see co-facts during the workshop but I would just like to highlight their work. This is a civil society thought that is in an end-to-end encrypted channel called Line that is very popular here, it's like WhatsApp. And I assure you that this is not gender biased if you refresh, you just switch to a different pronoun. Anyway, so in any case, the idea very simply put is that if you add this thought as your friend instant messaging, whenever you see a piece of deep fake or shallow fake, you can always share to this thought, which is literally called is it true or not? So it gives a good instant response that doesn't cost anything from one's psychology. If you see something that you suspect is this information, to fact check it, of course it takes far more time but it's very easy to just instinctively reply publicly, oh, is it true or not? And then it gets everybody into this like turn-based game instead of real-time strategy piece of mind and that people can actually look at it collectively and see what is trending in those. And they're all replied, meaning that the civil society is working very quickly, but you can see all the trending disinformation campaigns and rumors of the instant messages. And this literally shines a light on those dark places where it's impossible for search engines or other analysis to go through because it's end-to-end encrypted. But just by getting people to have it of reporting new screens of virus of the mind here, as we can see, this is almost always about elections now but still, there's something about food, there's something about diplomacy, any election, those three things could be a theme here but we don't share it for too long unless it pollutes everybody's mind but we can also see that the crowdsourced investigators are very quickly clarifying the messages. And the great thing is that there's some also great innovations like maintaining glucose levels by this new innovation by the E-Tree by our Institute of Research but it seems too good to be true so people respond saying maybe this is a rumor that tries to boost E-Tree's research prowess but then people start fact-checking and discover this is actually true, this is a breakthrough in medicine and then people are able to classify this as truth and also we learn something about if you package truth in this way, it actually goes viral. Again, this goes into the, if you shine a light a conversation that's going underground, you can review not only the falsehood but also the truth that this was sharing and so I encourage people who are interested in using machine learning, there's a lot of machine learning experts in the COVAX community to engage with the local COVAX community as well as other communities working on this topic because this is a highly interesting topic and far as I know they are also working on porting this to WhatsApp and other forms of instant messaging. Dan Biaz would like to know, what are the channels by which the PRC spreads this information through Taiwan's society? Well, there's many, like if there's any channel that you can think of, it's probably already used for that purpose but then any channel that can spread information is a channel that can spread this information. As I mentioned in my main talk, this information brings out of an environment where there's already distrust, there was there's already a discord in the civil society so that people are more willing to believe those conspiracy theories. So if we work in a misinformation layer, if we see those well-intentioned speculations and just occasional individual trolls and treat them seriously, authentic, then in that particular channel, people become vaccinated against future disinformation campaigns and so far that is the most effective way that we've discovered. It is just through deliberation. It could be an online AI-moderated conversation that asks people what they feel about any particular thing. I'm going to share an example. So this is the AI-powered conversation in System Polis that we have deployed since 2015. At that time, the first was to talk about this thing called sharing economy which means very different things to many different people. And so people just look at one particular example which is a Uber's use of people without professional driver's license and we just ask people one simple question, what do you feel about it? And then people share their feelings and see their feelings among the different groups of people who are their Facebook or Twitter friends. But we always start with a crowd-sourced fact evidence-gathering campaign where people contribute to not just open data but also independent reports and things like that that gives people a timeline and the basic facts of which to discuss. And then after that, we ask for people's feelings. And once people share their feelings, they get into this state of mind, what we call an overview effect where they can see everybody's feelings, not just their own personal feelings. And then we start to ideate. The best ideas are the one that address the most people's feelings. And so user experience-wise is very simple. You log in or you can remain anonymous and then you see an avatar representing you and then you see one feeling from a fellow citizen and then you can click agree or disagree. And as you click agree or disagree, your avatar will move among the clusters of people that you know or you don't know and you learn where you signed, where your feelings resonate with the most number of people. But in this space, there's two advantages. The first one is that there is no reply button. So the troll has no place to troll. It is impossible to reply to each other's feelings. You can only contribute your own feelings. And the second thing is that it shows that people's feelings can change when people propose something that is resonating with everybody. And so in a place that is safe, it's impossible to troll. And also it is strictly additive. People can only add to the feelings that resonate. We see something that is very different from traditional social media. People would agree to disagree on a few key things, usually ideological things, but they spend far more time proposing consensus statement that they wish to compose and in a nuanced, eclectic way and convince people in different groups because we share the agenda-setting process. This is like a more complex version of Slido that you're just using. You can add to Slido but you cannot reply to each other's questions. You can upvote and I commit myself of answering to the ones that has more resonance with people meaning more people want to hear the answer to that question. So through this visualization, we know time again and again and we work with like this is the following green consultation worked with international communities to perfect this kind of feeling gathering, feeling checking AI-powered conversation so that by the end of it people understand people have legitimate different authentic feelings and in this kind of place of mind it's impossible for this information to claim people's attention because people have already have a whole picture contextual understanding of what information there is and most importantly what a society feels collectively about this and in that kind of situation there's no way for disinformation to spread. So again, active participation in the stage where things are just rumors and misinformation is key from to prevent disinformation to spread. An anonymous person would like to ask how do we ensure the credibility of fact-checkers? This is a great question. Usually if they review how they do the fact-checking if you can do fact-checking on their method of fact-checking if there's an accountability trail that they can account for the kind of work that they do this is how we ensure that people can always check the fact-checkers. So the Taiwan Fact-checking Center which publishes reports of people who work harder are most likely to get fat for some reason. Right? And it's partially true. And they revealed exactly what are the main contentious points which the fact-checkers have consulted what kind of correlations instead of causations are there implied in the papers. And basically it says it's only if you work during dinner and if work distracts you from dinner and that there's maybe a causative relationship but otherwise the misleading or oversimplified title is actually partially wrong. And so of course everybody can see the research methods, the fact-checking method that the fact-checkers have used to clarify this message and it contributes to the public discourse so of course you can replicate their steps but also supply more materials and more evidences to this fact-checking process. So at the end I think everybody can participate in the fact-checking. They're not just consumers of reports of fact-checking but it enables in a kind of standard operation procedure kind of way how to approach a piece of news and how to approach information. Four people would like to know how is misinformation defined and what's the difference between misinformation and different perspectives. So misinformation to me personally is unintentional, certainly not organized speculation or piece of information that may contain controversies or disputes that are just partial truths. And in Mandarin here we use the term which means literally controversial information and this is the most neutral term that we can use because we understand that it's like a puzzle. Everybody has a incomplete piece to everybody else it looks like misinformation because everybody has a different slice of truth from their perspective. But the important thing is that we see misinformation as an invitation for people to complete the picture together the evidence and the feelings together and so this does not carry the connotation that people who are different from say the governments or the minister's perspective are misinformation. I can also hold many information that are outdated that my browser cache really need to be refreshed and things like that. It is outdated beliefs and outdated facts are actually by itself misinformation but not through malicious and certainly not through organized action to sow this court. And so this moniker translated as reminds us to be humble to people of different perspectives and always see that as an invitation of a real and true public conversation. Just would like to know which platforms are the most dangerous distribution channels of the F word that I don't use in Taiwan. I think this is a great question. Dangerous to whom though we see PTT for example being a mostly text only forum this is like the equivalent of Reddit in Taiwan it offers a very I would say very good tradition of moderators in forums individual forums in the PTT they have the right to self moderate based on the contributions they also run elections for their moderators and things like that. It is very much a kind of autonomous pooling public discussion forum that is free from any for-profit motives and so because of that but precisely because of that we see the most advanced disinformation campaigns and even criminal offenses first testing their waters on PTT because the entry barrier is the lowest and that people are often given the benefit of doubt testing new bots or new artificial intelligence techniques on PTT and so we see that it's where innovation happens but it's also where the newest threat happens is like event gosh compared to Facebook and to Line and Facebook usually serves as the amplifier of the techniques that's already discovered as useful on PTT so it's a disseminator and Line usually intentional disinformation campaigns that they know that they will not survive public scrutiny if it's posted on PTT or Facebook publicly but they have to rely on end-to-end encryption anyway encryption don't get me wrong it's a great thing but because of end-to-end encryption we need to do more innovations like the line bot and things like that to surface those disinformation trends into the public scrutiny so it's dangerous in various different ways but it's also exciting and innovating in different ways and six people would like me to elaborate more about the media literacy curriculum in K-12 what are the persons involved in planning the curriculum that's a great question so in the new curriculum that's going to take effect next September I believe there's a very kind of paradigm shift of how we approach education previously in our previous curriculum which is part of the course in many East Asian countries we emphasize important skills in perfecting the skills of various different disciplines but in the new curriculum the new educational target is characters or literacies and so the characters of autonomy of people's ability to design their own curriculum essentially is the first thing that we want to cultivate and second interaction and media literacy is part of the computational thinking and media literacy character and the third one is common good that is to say see people with different cultures and different ethnicities different backgrounds not as instruments but as people and collaborate share the values that are of use of everyone instead of using other people as means and so these are the three fundamental characters we want to share with our new curriculum and that costs for a very different kind of curriculum design instead of individual schools just executing what a ministry of education and the curriculum committee have designed for them they are actually now co-creators with the parents with the community and the students and so we have a new set of capstone projects especially senior high school level where the students solve a social or environmental problem through say social entrepreneurship or any other contribution to the sustainable developments as part of their learning and they designed the course with the local NGOs and non-profits and social entrepreneurs and that's the senior high level and the senior high students can choose the ones to their liking so we bring the college system the college credit system to the senior high level and to the primary and the junior high level we also encourage different schools to emphasize on different literacy projects by again engaging the alternative schools, the experimental schools the local college universities and so on and use those different modes of education as resources and co-create across generational barriers and so that the children learn to contribute to society by those literacy campaigns and joining those volunteer efforts and maybe becoming a fact checker as part of their curriculum and again this is devolved to the individual schools curriculum planning committee and not at all determined by the national academy of education research but we switched our evaluation system our examination system to look at what kind of characters of collaboration that the students have born out of those capstone projects that they have finished or completed and how they see their own social their positive impact to the society after completing those courses and so this is the kind of overview but there's much more to it and if you're interested feel free to look at the NAER the national academy of education research which we have a very complete material about the new curriculum and how we partner with our NGO county parks for people who like to know how do we prevent influential politicians heads of state and state actors from using this information to misguide their own people like these people are seeing in the US well first of all I would like to say that sometimes people use Twitter the way Twitter is meant to be used that is to say a self-contained short message that people cannot misinterpret and I actually admire these kind of views of Twitter but that said I understand also that people sometimes take those as cues and then spread far more malicious and trolling information that may not be intentional by the original state actors that's hosted on Twitter I mean I post on Twitter all the time and people partially quote me and make various mess out of those messages so it is not about prevention this is about actively engaging this is about asking more transparency and more accountability in the policy making process like for me for example personally I publish on my website the public digital innovation space all the raw transcript of all the even internal meetings but also meetings with media and also with lobbyists in a complete transcript kind of way so this gives people what is they like to be in my day-to-day work and gives me a great way to respond to any online accusations or partially quoted information eventually just linking to each word in every word which all has its own permanent link it also makes it very easy to be discovered by the search engine and this is a great piece of civic technology built by my society folks in the UK called Say It and by personal experience this really empowers independent journalists and investigative journalists because by sharing the context of policy making not just the end result of the policy itself but why and how of policy making it empowers the independent investigative journalists to add their perspectives to the narrative and also encourages people to look into the various different complexities during the policy making process instead of just relying on one press release because it's just a compressed version in a tweet I mean it's just 140 Chinese characters or 280 English letters it is insufficient room to give the total context but in my Twitter post I often just link back to the Say It which I mean a URL it's shortened doesn't consume much letters but people get into the context of seeing one very poetic provocation from me but it is backed by a full conversational context so I would encourage other Say Actors to do the same so Jinhao and four other people who would like to know is there a sharing and mechanism and partnership established for the government especially the administration to share what we know about those disinformation campaigns yes so as many of you know in our administration's homepage there is already a real-time clarification where all the different ministries respond usually within five or six hours whenever there's either disinformation or just rumors misinformation or just difference in perspectives or criminal offense and they are actually really quick nowadays to respond to it so much so that when people see something on the morning news and wonder is this true or not people generally learn to wait until noon and then there is a new batch of information shared by the central administration but we also hear from civil society and journalist partners that this display confuses misinformation disinformation and criminal offense sometime we publish something that the ministry beam as intimidating the public and has some origination information that they share and they actually brought up lawsuits and investigative processes but then it leads us to think that all of them are here are like this but it's not like that only maybe less than 10% are criminal offenses that the administration shares most of it are actually just complimenting each other's information and preemptively revealing more context in order to respond to would be disinformation campaigns and so we hear from the civil society and the journalists that they want more clearly coded maybe collars maybe icons clarifications they want to know which one of these are criminal offenses which are the disinformation campaign that we have some idea of and which are just our conversation with the media and the civic media and it also makes it much more clear that is we're not targeting journalists we're not offending journalists this way if we're sharing the sufficient contextual information and clearly marking it and so in response to jihad we're now working with the various departments in the administration to provide the right category in right coding of that bulletin board but the revised version and mechanism will still be on the same webpage on that bulletin board and so that's all the question I have the time to answer because we're in 10 a.m. now but feel free to look at each other's questions because those will inform the workshops and I wish you a great training session in workshop and let's collaborate and bring the light into democracy thank you