 Thank you director and chairman son So we will now start our very exciting panel discussion the panel discussion will be moderated by Professor and dr. Kaori Hayashi of the Institute for AI and beyond and Of graduate school of interdisciplinary information studies of the University of Tokyo, so professor Hayashi, please Hello everybody, and my name is Kaori Hayashi of the University of Tokyo And it is my honor to be moderating this exciting panel discussion today And let me begin by introducing our three exciting panelists first digital minister Audrey Tan of Taiwan Hello, welcome to our symposium Minister hello We are all hello. We have all very grateful and excited to have you join this important occasion for us Minister Tan is famous in Japan and the world for her work Among other things in Taiwan's fight against the novel coronavirus through the use of digital technologies Minister Tan is also well known in Japan for her books such as the future of digital innovation or digital to AI no mirai o kataru our next speaker We also have Sputnikko who is associate professor in the department of design at Tokyo University of the arts And Sputnikko is an artist designer who is known her for her film and multimedia installation works that explore the social and ethical implications of emerging technologies Her works have been featured in such venues as the Milan international design Triennale and Mori Art Museum and Last but not least is Dr. Yoshihou Ikeuchi Associate professor at the Institute of industrial science of the University of Tokyo and a member of the Institute of AI and beyond Dr. Ikeuchi heads biomolecular and cellular engineering lab which studies mechanisms for new neuronal development and related brain disorders Using unique in vitro systems So before starting the panel discussion, I would like to ask each panelist to give a brief Initial statement about their interests in and concerns for AI So first Minister Tan, please give us your statement very happy to be here and I will start by reciting my job description as Taiwan's digital minister it goes like this When we see the Internet of things, let's make it a Internet of beings When we see virtual reality, let's make it a shared reality When we see machine learning, let's make it collaborative learning When we see user experience, let's make it about human experience And whenever we hear that a singularity is near Let's always remember the plurality is here the plurality enables us to work on everyone's problems with everyone's help And that's the idea of collective intelligence During the pandemic, there was in Taiwan, and still are in Taiwan many availability maps for the mask, the PPEs first sold in pharmacies and then at convenience stores But it's not an idea by the government It's not even an idea by the business sector It's by civic technologists In the G0V or GOV0 community there were more than 100 different maps all over Taiwan catering to, for example, the elderly people who are more willing to use chatbots the more young people who are OK with automated navigation to the pharmacy that still have those masks available even to people who are not very old and like very, very young and in a language that they can understand, for example through this very cute Shiba Inu spoke stock that caused all the young people to spread the idea that if you wear a mask, you protect yourself from your own unwashed hand And the result of trusting the citizens with real-time open data enabled each and every one of us to think of better ways to distribute the PPEs, to use it better and also disseminate this knowledge That's collective intelligence And when the collective intelligence shows tensions in the society when people have different positions, that's where AI can help In Taiwan, we have already institutionalized the use of a system called PULIS which is AI-powered conversation What you are looking at is a real conversation in 2015 when people talked about the UberX phenomenon of people driving without professional driver's license, strangers and charging them for it But instead of fighting among the polarized ideas on the more anti-social corners of social media AI, in particular, this social sector-operated AI called PULIS showed everyone how much of a resonance can people bring from our initial positions So where there are, of course, a lot of ideological differences The AI, through the use of clustering algorithms, showed people actually we have much more common in ourselves with our neighbors on important issues like insurance, registration and so on on the UberX phenomenon and we very quickly put that into a new law and Uber worked with multi-purpose taxis in 2017 and this shows us if we deploy AI in a way that assists our collective understanding then it's like my eyeglass, it helps me see but it doesn't compete with me in seeing It is accountable in the sense that if it is broken I can fix it myself or anyone can fix it and once it's aligned and accountable this way then it adapts itself to the societal needs empowering everyone in the society without resorting to what I call authoritarian intelligence which is a central planning that deprives the society of the democratic control So that is what I mean by assistive collective intelligence and I look forward to have the panel discussion with the two panelists Thank you Thank you, Minister Tanda, it was a very wonderful and inspiring statement Next, I would like to ask Sputniko to give your remark Okay, thank you I'm going to go through my slides Oh, hi everyone, I'm very happy to be here and I'd like to first introduce who I am and what I do Can the slides go on, please? Can you go next, please? So I'm an artist but actually I have a computer science and mathematics background Could you go up next, please? So I studied maths and computer science at the Imperial College London and after graduating, could you go next, please? I studied design in the Royal College of Art in London and after studying design, could you go next, please? I became an assistant professor at the MIT Media Lab So I taught at the MIT for four years and I had my own research group called Design Fiction Group Could you go next, please? And then after that, 2017, I came back to Japan after a few years and I became a project associate professor at the University of Tokyo in student industrial science Could you go next, please? And currently I am an associate professor at the Tokyo University of Arts in the Department of Design Could you go next, please? And other than teaching and also showing, creating and making works I'm also giving talks, lectures For example, I became a TED Fellow from 2019 I'm also a young global leader at the World Economic Forum from 2017 Where I gave a talk in Davos, thank you And my main research theme is called speculative design And that's this idea of design which was first coined by Professor Anthony Dunn and Fiona Rabie in RCA in London And this is a design that asks questions about the future So this is a design to stimulate discussions about the social, cultural and ethical implications of technologies Could you go next, please? So one project I did in this speculative design theme is that this is a project called menstruation machine So I designed a machine that allows people to experience the whole process of menstruation And the reason why I designed this machine is that half the population of this planet goes through menstruation Bleeding, pain, all this really nuisance But I was really concerned with the fact that menstruation in society is often treated as taboo Something that we shouldn't be talking about, something that should be hidden And I was really frustrated that it's affecting half the population of the planet It's giving so much trouble, but why can't we talk about this? And why can't we talk about this to think of ways to make it easier, make it better? So that's why I designed this machine that allows everyone to experience this process and talk about it Could you go next, please? So I created the machine and I also made a video which is on YouTube And when I posted this video on YouTube, it became really viral And then it ended up being exhibited in New York MoMA or Museum of Contemporary Art in Tokyo This is a show in Germany So this is the kind of way I'm working that I do I design about technology and I try to ask questions about the future Could you go next, please? So one issue I like to bring up in this panel discussion is that although AI is amazing, it's bringing so many amazing progress, development There's some really concerning topics about the AI And one issue I like to bring up is this issue of bias So artificial intelligence, basically the aspect is basically it learns from the present and the past data You know, machine learning is all about learning from data which means that it also could learn from the biases that exist in present society So this is a scandal that came up in 2018 And Amazon, they were trying to develop AI for hiring But actually Amazon in the past, they didn't hire that many women They were very male dominated hiring So the AI actually learned from the past data that okay, maybe it's a bad idea to hire women So Amazon found that this AI tool was actually discriminating against women when the resume had anything to do with female or women So Amazon luckily discovered this bias and they stopped using this AI tool But I think it's a lucky thing that they discovered this And I'm very scared that because AI is such a black box And especially in Japan when gender gap is not very great in Japan I'm not sure if many people would discover and realise this kind of bias existing in AI Could you go next please? I'm sorry, I'm a bit overdone Another issue is that similar things is happening with racial bias So this is a scandal that came out where in the United States they're using this AI called Compass to analyse the probability of a criminal committing offence again And they found that this AI is giving a harsher verdict to people of colour So this AI was thinking that black people has a higher probability of committing offence again Which really affects the amount of prison sentence, the bail that black people get compared to white people and this is a big issue Could you go to the next slide? Sorry to go over time but the big thing I'd like to really ask today is that Okay, AI is great but how can we make AI bias free? And how can we make AI that solves many issues that we have in the world of inequality Gender bias, racial bias and also things like climate change or sustainability And I really hope that Beyond AI can become an institution that really thinks about how to solve these difficult issues we have today Thank you very much, sorry for the time Thank you Thank you, Sputnik, it was very inspiring and also provocative statement You brought up some sort of dark side of the technologies And yes, we will discuss that later And finally, Dr Ike, would you please? Thank you very much Thank you very much, I'm so excited to be here And also I'm really excited to be part of Institute of AI and Beyond to do research Could I have the slide please? So let me introduce myself briefly I summarized my resume in three lines First, I was doing PhD student work on genetic code and protein synthesis Which is really fundamental process and machinery for all the life on earth And it is very interesting And it uses a lot of digitalness and analogness to make the life able Be able to live And it's really interesting, but I will not talk about in detail But I realized that very fundamental machinery, even very fundamental machinery Is repurposed in different ways And such as forming neurons, shape and connections So therefore I moved my interest to this shape of neurons And did my post research in the United States But then I came back to the Institute of Industrial Science in the University of Tokyo And then I started doing this new work on making neural tissues from human IPS cells Can you go next? So our brain is, you know, as you know We have different regions within the brains and they execute different functions And very interesting thing is that those different regions are connected And then they can function And this is really interesting and unique part of our brain But not similar in any other tissue or organs in the body And in biology as AI is growing and it is used and it's very exciting But at the same time, also stem cell technologies are really evolving And that is making it really exciting time now And as you probably know, human IPS cells can be generated from cells in our body And which is called reprogramming And that reprogrammed cells can become anything, any cell type in our body Including neurons, neurons are the cells that makes our brain And they're connected and then they process information Not only that, we can make three dimensional mini tissues And it's something called mini brains and we call it organoids in the field And those mini tissue can be generated and including brain mimicking tissues But as I mentioned before, brain is really connected and that makes it function But the current technologies of organoid for brain Is to make a single region or a few regions separately So we decided to make a way to connect them each other in a tissue culture chip That is shown here, which is developed together with Fuji Sensei And by that, as shown in the right bottom in the black dark background You can see that the mini tissues that are about one millimeter Is connected each other by their own extensions that are called axons And we think that it's mimicking connections within the brain Could you go next? What we demonstrate with this is that if you didn't connect, which is shown on left You can see the activity of neurons are quite periodic, simple And not so complicated But if you connect it, we're quite surprised But they really show complex and intense activity And actually, we think that they can memorize some sort of simple form of information For short term So we would like to make it long term memory And we would like to know why they can really memorize some sort of simple information So with the Institute for AI and beyond We would like to understand more by asking three questions below Can we know more about circuits in the brain from this in vitro system? We also want to ask questions, can we make them function with the help of AI? And also, this is very exciting to be in the group of other people who studies AI So we would like to ask the last question of can we or can the research inspire AI development? And we're doing this with very talented research associated Tatsuya Osaki And we would like to recruit more people I would like to introduce the last slide of, can you go next? With the little bit of interaction with society So as Spock Nicholson mentioned that RCA does a lot of speculative design And it's really well known for it And the Institute, I mean IIS and RCA set up a new lab that's called Design Lab Which facilitates the interaction between scientists and designers and students all together at the campus And so they form small projects And then they try to come up with the usage of our technologies into deployable products And here I'm showing three different examples Which I'm not going to be able to go into details But through the interaction with the non-scientist people or intermediate people We can know what people really want from a technology or what is lacking at the current technology or society And through the interaction we're so inspired and we can apply those ideas into the research And we can accelerate the research and we hope that will be helpful for the society And that's it from me, thank you Thank you So thank you Dr. Takeuchi for your very fascinating presentation Now we move on to our panel discussion And may I ask Minister Tan first to kick off our discussion by offering your responses to the statements? Certainly First of all I think about the bias issue all the time for being a non-binary transgender I'm always being biased against Very seldomly biased for anyway And so this I feel very clearly And I think about the mask availability map about how it looks fair when I first see the map on my phone It looks like the pharmacy distribution is almost perfectly lined up with the population centers So I feel very good about it But just a week after a member of the parliament VP of data analytics of Foxconn before she went into the parliament MP Gao Hong-en interpolated our Minister Chen Shizhong saying You didn't take into account the public transportation times that the rural people need to take to reach its near pharmacy So even though it looks like the same distance on the map It's not the same time in terms of the cost or opportunity cost for the rural people So it's actually biased against them And because we share the open API updated every 30 seconds Minister Chen Shizhong our Minister of Health were able to simply say Legislator teach us And I think this quote is very important because it shows that in an evidence based culture Any bias as soon as it's discovered have nowhere to hide The bias may already be as Sputnikko put it in the mind of the planners But the AI and its visualizations if democratically controlled Allow the society to have this sort of feedback And it's always constructive in the sense that every MP or every civic technologist can offer a better way To organize the distribution and we did fix that just 24 hours afterward So such a feedback system I think is paramount if we are to use AI as assistive intelligence Thank you So I think the word if democratically controlled is something that we can dig into deeper So Sputnikko would you have something to say to me? Yeah I agree with Minister Tang about the feedback system But sometimes I often worry that the commercial incentive can be extremely strong Especially in technology So to develop to the desire to develop something very efficient Something very fast is very strong And I think we really need to have a strong force How can we build a feedback system that can be checking on this? Because we like to know maybe it's a good way to think about a tool where normal people could suggest biases Or evidences that where AI is working against them But often times you need to be pretty educated to be able to spot these evidences So I sometimes think like maybe a research institution needs to really fund that effort To check on this really fast development of AI and capitalism I don't know I wonder what's your thought on that How do we check and how do we create a sustainable feedback loop of tech and capitalism and checking where it's going? Yeah I think our current forms of democracy which was designed for an analog system And at most radio and television system really prioritizes this dissemination But not so much the feedback The bit rate of democracy if it's constrained only to voting It's just maybe four bits every person every four years And that's very very small amount of bit rate in terms of democratic control So my work on collective intelligence is almost always to increase the bit rate So that anyone who see anything wrong with the AI that's deployed in the society Can just go to our national participation platform, join the GOV, the TW Which has more than 10 million people using it so half of the country mostly And we have more than one quarter of petitions by the people who are not even 18 years old So the people who are learning this is not thinking about pay I have to wait until I'm at the age of vote Or when I'm digitally literate or media literate Rather their civics class teacher shows them by giving them assignments on starting popular petitions Saying that in order to understand data stewardship in order to understand bias There's no escape from operating your own for example air box that measures air quality That rides into a distributed ledger and then data stewardship will make sense Which is why we don't call it literacy anymore, we call it competence Literacy is when you're listening to radio or watching television Competence is when you are a producer and actively setting the agenda for the society Which is day to day not once every four years Yes, thank you very much So my question would be that how would you nurture this kind of competence Or what is your thought, would you give us some thoughts about what the ideal education for the kids Is to be able to empower them to be able to participate in such a democratic system Yeah, first of all we heavily prioritize free software in a sense of freedom in our basic education system This is not just to avoid vendor lock-in But it's also to make sure that they are empowered instead of feeling a sense of helplessness When the machine's algorithms turn against them Indeed if something is free software it means that every student is free to fork Meaning taking it to another direction without violating copyright or other so-called intellectual property laws And this is I think number one It's very important to have the agency over the software and think in terms of code Not in terms of the algorithm It's like the central planner's way of thinking about society and code is something that everyone can change And that's the first thing And the second thing I think also equally important is to make sure that when the society detects such a bias The discussion takes place in a public infrastructure In for example real world we have a conversation like in this panel in an academic setting Or in a public park in a public library in a town hall These are all public infrastructures in the commons But in the online world if we do not have the sort of participation infrastructure like the joint platform Peril is PTT which are open source and social sector operated Then we're confined to the economic sector capitalism surveillance capitalism Operated corners more like nightclubs and pubs and drinking places Which do not actually foster public deliberation They may have private bouncers and sell you addictive drinks and it has its place But that place is not public deliberation about the results or the loss of agency to central planning AI algorithm So have a public infrastructure intentional invested by the social sector and the public sector is also equally important Very interesting The strengthening of the free software availability and the public infrastructure are the suggestions Are you persuaded? I have a question about the Polis tool that you just mentioned Because communication is another big issue I'm interested in Because communication these days like fake news or extremism is much easier to spread online And you just showed just now in the presentation you try to design a tool That where AI is used to try to find similarities between people with opposing opinions Is that what you said? I'd love to know more about this because Sure, right, the design is very simple You see your friends and families not nameless trolls on the system It's by payments clustering so people who feel similar are grouped together And the main contention points discovered via principal component analysis are projected into two dimensions that are dynamically updated So the way we input into the system is by looking at one statement 140 characters So tweet length and for example this was my contribution in 2015 I said I think passenger liability insurance is very important It's my personal feeling there's no right or wrong but you may agree or disagree If you do agree you move toward me if you disagree you move farther away from me It's like a open space technology but there is no reply button When there is no reply button there is no room for troll to grow It doesn't incentivize trolling and before long people discover that we actually agree with most of our neighbors Most of the time on most of the issues simply because we simply agree to disagree on those ideological differences Because they do not sell addiction This is a free software platform that prioritizes a more nuanced understanding come of feelings And by focusing on feelings not interpretations it makes us more pro-social How do you feel about the design of the current major communication platforms like Twitter, Facebook And what happened recently with the United States capital riot where it seemed like there were many conspiracy theorists like QNHON Spreading online and it was creating a lot of Havoc online How do you feel about the current design and what are the improvements that could be done? In Taiwan we have this idea called humor over rumor And indeed the cute Shiba Inu is part of the campaign to make sure that whenever there is a trending misinformation People just reported much like we flag spam from our inbox And the situation of the current generation of more anti-social corners of social media It's like an inbox that's almost all spam and just a little bit of it's meaningful communication And that's of course not tenable But I remember that because I guess old enough to remember Bill Gates also used to say Email will be broken very quickly if we do not have the postal stamp and charging them for it But it's not soft like that not through commercial incentives But rather through people volunteering to flag things as spam And for international networks like Spamhouse to signal a untrustworthy spammer So that its email moves to the junk folder rather than to the inbox And so similarly in Taiwan we have an immune system from the social sector Where people just crowd checks like through the COFAX initiative Crowd checks anything that is trending or about to trend And make sure that if it has a high R value then the response to that Which is even more fun has even higher R value so it inoculates people who laugh about it And I think this is one of the ways to solve it I'm not saying it's perfect but it beats censorship in my opinion It's very interesting So the application of technology is a key to our discussion I think So maybe some thoughts from Dr. Ikeuchi How as a scientist do you think to for example your science, your experiments What do you think is the appropriate use of your technology that you are now doing research on? Our technology So I have to be honest about our technology is really far away from actually deployable product So as I mentioned last part of the slide I showed that we tried to interact with design people To understand what is the need from the society, what is the problem And we would try to of course come up with a scientific solution In our case a neuronal tissue or neuronal computer or the computer mimics our brain That are made of cells Which is developed which we think is a trajectory from our current research And we would not be able to really sort of directly have the solution for that question I'm sorry but from the cells But we would be able to actually bring up a lot of thoughts and ideas and problems through that course For example so if we make the brain like tissue That might be actually as biased as our brain For example like that might be a problem for example But then we can have that type of discussion with the participants And we would be sort of trying to addressing those questions by of course by developing our thoughts But also try to incorporate others But the short question in the end is what technology cannot be used directly right now But we are quite enjoying the discussion through the design lab You are inspiring AI technologists Yeah I hope that's the case So I hope that is by directional So of course we're not really talking with the general public We're maybe in the small clusters that are close together in the PCA analysis But anyways but maybe we can have that little window from the society to have the thoughts and knowledge And we would be like to addressing more And I hope that's happening more in this institute as well Yes thank you very much The whole symposium is about how we best use AI technologies And what can we do about education and also about research and ethics and so on But we have to pause our exciting discussion now I have many more questions from our colleagues for Minister Tan And I would like to ask because of the time one personal question to Minister Tan and one more conceptual one Is it okay Minister Tan? Yes And let me start with the personal Why did you change your career at the age of 33? And what drove you to become a political activist instead of staying as a tech entrepreneur That has something to do with also how scientists, what kind of role scientists play in society And what kind of role political activists play in society And how these two are related together Would you respond to this question? Sure In 2014 March I participated in the sunflower movement Where we occupied the Taiwan's parliament for three weeks And the occupy was a demonstration but it's not just a protest demonstration It's a demonstration in the sense of a demo like a software demo And the demonstration was a very simple hypothesis That the collective intelligence of half a million people on the street And many more online co-organized by more than 20 NGOs Each specializing say in the human rights, in the cybersecurity implications Labor laws, environment and so on about a trade deal Can actually deliberate such a trade deal even better than the parliamentarians could do In their formalized hearings And the demonstration was a resounding success We showed that with good open space technology and nonviolent communication People actually understood the repercussions of such a trade deal If empowered with the kind of collective intelligence tools That shows them exactly how such a trade deal affects them And this leads to high quality conversations And make the democracy, well I guess more democratic And it was endorsed by the head of the parliament at the time And we had the national forum And we collectively decided that we didn't want the parliament to get occupied whenever there is a controversial issue And we want instead to take such a conversation to the digital realm And anticipate such controversial values And so that we can settle on what we call a rough consensus through such exploratory design tools Not last minute occupies So that's the kind of outside game that brings me into the world of applied collective intelligence In the government when I was recruited as a reverse mentor as one of the ministers at the end of 2014 So you are a political entrepreneur That's right, I'm a politician, you can say that Great, yes And I have a privilege to ask one more question It is rather a conceptual question And how can AI, this is a question from all of us All the panelists here, three of us How can AI help us integrate divisive society Or in other words how can AI resolve this society's deepening rift of inequality instead of strengthening it And I would like you to address this question by talking to our students what we can do about it Yeah, I think one of the main promises of internet technology is that it's not just for downloading It's not just for listening to mass communication programs that were already fixed by the programmers But rather everyone is invited to be not just a programmer to invent your program But also a civic hacker in the sense that you can take such programs and adapt it to your local need However you feel appropriate without the explicit permission or even imagination by the original designers So they become more like Lego blocks Now it's arguable how much internet has fulfilled its original design spec But that spirit is always there, it's called end-to-end permissionless innovation So my suggestion is that to think about AI as like fire, which is also a technology that can burn cities It's very destructive, we're not used correctly But we don't restrict the use of fire to just a few elites Instead we teach about safety of fire and safe fire use And applied batch processing of food through fire as early as maybe six years old In the kindergarten it's called cooking classes And so think of it this way about how to cook for the better And also how to share your recipe with others Thank you, that's a very inspiring comment Lastly, maybe we can have a comment from our panelist, there's Wu Niko and Gio Chi Well it's been great to talk with Minister Tang and you guys I hope I can see you in person in Taiwan or Tokyo after the Covid crisis But I think another key is equality in education is so important Open education for people of all class, all gender or race And I just hope more women should be studying engineering and science Especially University of Tokyo needs to work hard on that I agree And I think more like quality in education leads to better AI, better design for AI So it's been great to talk today, thank you so much It's been really pleasure and really fun and I learned a lot today too And it's really difficult for us to interact on this Covid situation Including this institute has been doing a lot online But it is a good thing that we can connect even to Taiwan It's really nice time also And we really want to cause more technological intense fire from this institute And then hopefully that can be applied to the society more in the future Thank you, and Minister Tang final comment from you? I wish you all live long and prosper Okay, well unfortunately we have just run out of time And I'm sure more can and should be said but let's save this for the future discussions And I would like to thank each panelist and particularly Minister Tang For joining us from Taiwan, Amidst your busy schedule Thank you very much And it was great honor for me to moderate this session Thank you, thank you everybody Thank you, thank you so much, thank you so much Thank you, thank you, thank you Now back to Yu-jin Thank you very much, wow that was really exciting I wish it could last longer but we'll just save it for the next time Thank you once again Minister Tang Spotniko, Professor Ikeuchi and Professor Hayashi For the wonderful and inspiring discussion We'll take a break now until 11.10 am, just about 10 minutes And resume at 11.10 am To introduce the research activities of the Institute for AI and beyond Thank you