Tembusu Conversations with Dr Lee Kai-Fu

may I first request that you switch your electronic devices to silent mode please thank you may I now call upon associate professor Ho Chi Kong master of tambusu college to first introduce the tembusu conversations as well as our distinguished guests for tonight Dr Lee profo please thank you good evening Mr Slim honor speaker Dr Lee kaifu students and college fellows I'm Ho Chi Kang Master of the college and also the vice dean of the youngster conceptual music in charge of academic affairs it's been a while since we are in this room without our mask on so welcome to our maskless symposium [Applause] so I want to extend our very warm welcome to you you know to today's inaugural temperature conversations which is a flagship event initiated by Mr Slim our director under the invitation of Mr Slim the tembusu conversations brings you experts practitioners and areas of Innovative Technologies finance geopolitical and social shifts to share the exclusive insights and perspectives with our college now tonight's speaker is very special it's Dr Lee kaifu who wrote seminar book AI superpowers and the new world order in which he predicted China's rights to become an artificial intelligence superpower now since the publication of that book this year you know here's another new book and so I'm making a plug for a talk release a book um AI 2041 10 visions of our future where he shares further predictions on a set of Advanced Technologies that will greatly impact the world now this evening Dr Lee Rocher is insights on AI development deep technology and technological Revolutions in areas such as Quantum Computing annual energy and Life Sciences so without further Ado let us warmly welcome Dr Lee on stage uh thank you very much it's great to be the first speaker in this series and I want to talk to you about the amazing times in which you live and the opportunities afforded by the what I call the fourth Industrial Revolution so um in 1928 uh sorry in 2018 I predicted that China would become an AI superpower and that was quite controversial at the time and this is a graph from the book where my predictions were that China would Advance itself in a number of areas in artificial intelligence and today it looks like that has happened so hopefully my next book will be equally accurate in in its predictions and if you look at AI maybe just a few words on what is AI when I say AI I'm talking about narrow machine intelligence that collects a large amount of data in a particular domain and becomes incredibly good in optimizing some objective function so if you were an e-commerce provider you would recommend things that the user is likely to buy with the objective function being maximizing your revenue and if you were a social network provider you would be maximizing user interaction and their likelihood of clicking and watching content and AI is so good at that because it collects data from hundreds of millions of people and learns your similarity to your similarity to them and based on your previous preferences it can predict at what kind of content you would like that's the internet type of AI of course a bank can use it to base on many people who borrow money and they can train a model of people who actually return the money versus a model of people who defaulted for new applicants if you're more like a defaulter they don't give you the money if you're more like a repairer they will give you the money and then this can advance itself to perception AI where AI can see and hear like people and it can move to autonomous automation AI autonomous AI where AI can manipulate and move and these are developed in parallel and we're now seeing all types all four types of AI really all around us and China became very strong in in AI for the reasons described in my first book had china has a fantastic engineering education as thus Singapore of course and China also has a large number of these engineers and today actually everywhere I go people talk about the shortage of AI Engineers but China has a surplus number two China has had tough internet entrepreneurs who in the market with the most users also build super apps that capture the hours of users each day so more users times more time means more data which then number four here shows you that the more data trains better models one of the key points I make in the book is that in the era of AI data is the new oil so the more data you have the better off you are and then of course China actively has a lot of funders of AI like my own company as well as supportive government policies China has about 25 AI unicorns and we funded about 10 of them and collectively they're worth 30 billion us and they spend all kinds of areas that you can see on this slide so hugely successful and today in the new book that I wrote last September I predict that China will further extend its technological success from AI to other new areas and they're listed here the five key areas are briefly go into each of them and talk about the opportunity that exists for for China and how it would grow but before I go into these five areas I want to re-emphasize what a special time you live in uh probably the closest approximation was 120 years ago when three great inventions changed the next 120 years and these were electricity electromagnetic wave and the combustible engine they created the automobile industry the and and the electricity electrical grid and the radio TV computers internet were all built from these revolutions and they not only created great companies like Ford and General Electric but they changed the landscape of technology and competition and everything is different how people today communicate play and work are completely different than 120 years ago largely enabled by these three inventions and their popularity made by practitioners in the United States we're at a similar opportunity point today and these are the 5 five technologies that I claim to be as important as electricity some of you may be skeptical so let me quickly go through them artificial intelligence as I said is a technology that can beat us in every single domain where we can get data within that domain I mean isn't that more important than electricity it is truly part of human intelligence what we considered impossible is now not just replicated but beaten by Ai and we see that everywhere number two automation Ai and Robotics will take over and replace at or enhance at least 50 percent of human jobs within the next 20 years so jobs will change Mark and and routine jobs will be gone and we as human race well for the first time be liberated from having to do routine work isn't that as big as electricity number three is quantum this is a new way of computing that will solve unsolvable problems on today's classical computers it will run trillions of times faster and solve problems related to security encryption climate modeling of human beings because it is based on the fundamental principle of quantum mechanics so it can accurately model anything that exists as life or in in on Earth so the ability to solve unsolvable problems I mean isn't that as big as electricity number four is life sciences where scientists will be able to create life and modify life in a way that's never been done before this is not medicine this is not treatment this is creating a species like a worm that can behave as a fertilizer it is like doing things like editing a human gene so that we can forever be immune from a certain disease isn't that as big as electricity and lastly new energy will bring down the cost of energy through new technologies like hydrogen and fusion and others that will reduce the cost of energy by 90 in the next 15 years probably more in the next 20 years making it like making energy almost free this is not about green of course it is green but it's all it's about low-cost energy and now collecting these five things together if you think about the cost of any good today your phone your glass your pen your chair it has three basic costs first the materials second is the laborer third is the energy that's the three major cost of making anything the cost of the material will go down dramatically as life science scientists work on building new materials one molecule at a time and not having to make them out of toxic and limited materials like fossil fuel that will bring down the costs the labor will be replaced by automation robots will make future products thereby dramatically reducing the marginal cost of making a product and the new energy as I said will go down by more than 90 percent so if all goods go down by 90 in costs we as a human race will have the fortune and the ability to make enough goods for everyone thereby potentially eradicating poverty and hunger from the earth and isn't that as big as electricity that's the opportunity that's ahead of us and it's in the next 20 years so you were born exactly in the right years to participate in this Revolution and also to enjoy the benefits from the revolution so in the next 15 minutes I'm going to spend three or four minutes on each one of these technology areas it will have to be very fast I'll speak at a very high level give you an idea but if you want more details buy my book so AI is beating humans in everything we can imagine I've talked about AI being an omni-use technology that can be applied everywhere from internet to business to be able to see and hear to being able to manipulate and move and and this will be a tremendous progression that will help us do white collar and blue color routine work that will liberate us from having to do them and two big examples is on the left side on the upper part we see the blue bars showing AI beating humans in the ability to recognize objects which created many applications and opportunities such as inspection of goods in manufacturing reading of radiology and diagnosis robots that can see and cars that can drive themselves and we at sanovation invested in many of these companies creating many unicorns on the bottom hard we see a more recent breakthrough where the blue bars again show the computer performance in reading comprehension that's measured as in feed a book to an AI algorithm ask any questions pertaining to the book AI will answer the questions better than most people and that has already happened if you didn't realize that you should look it up that is a huge impact because human language is arguably even more important than human Vision because it is the way we communicate and the way we record history and the way we record all human knowledge so the understanding of all human texts is phenomenal and that will lead to breakthroughs in speech recognition machine translation search engines will go from one query to a bunch of websites you can click on to one question and get one precise answer and targeted advertising this one is a little bit scary for the same products will be advertised differently to each individual based on that person's needs and preferences and obviously it has tremendous commercial opportunities but it also has many downsides because it is in some ways similar to what Cambridge analytica did but much more powerful this is Cambridge analytical on steroids both for the opportunity for doing good and for doing bad one of the breakthroughs I mentioned is called self-supervised learning it is that language thing I talked about at the bottom basically AI that gets better with data more data the better so what if we have infinite data Well we'd have the problem of how do you teach an AI with infinite data with trillions of pieces of data well the answer is you don't have to label the data you don't have to say this is a dog that's a cat you just show so much data and tell AI to self-organize its structure so that they can remember synthesize and generate data that it hasn't seen before so essentially it's reading every book in the history of human race and trying to predict the next sentence within one of those books having read everything before it and with the nearby context and when it can be compressed to have a abstract representation it can then be used to recognize or synthesize or answer questions leading to tremendous applications and and the system would work as scan and read everything in the world and have a conceptual knowledge in a large language model and that model can be fine-tuned to answer questions do targeted advertising recognize speech Translate Arabic to Chinese and so on and so forth this has already reached substantial progress one of the companies we invested in has made such a model and I was recently in Middle East telling them about this and they said well you don't know Arabic and in three weeks the machine learned Arabic without without anyone on the team who knows Arabic and it actually beats Google in the performance of translating from English to Arabic and this is all machine learning by itself on the tremendous amount of data so China is once again participating in this newest AI Revolution if you look at who invented this self-supervised learning it's American Research but who made it practical it may very well be Chinese similarly who invented the computer vision algorithms the ones that led to the bar Americans and Canadians and some British and but who made it practical is the Chinese Engineers going back to my point about China being really the engine of development execution and fast growth fueled by a tremendous engineering population and a very hungry and tenacious entrepreneurs with whom I had the fortune to work with moving to the second area about robotics there are many ways to do robotics you can apply it to Transportation that's autonomous vehicles applied to manufacturing that's AI for the industrial applications these are some companies we invest in and the approach we take is very different again from American companies the American Automotive autonomous driving is characterized by waymo which basically tries to build a better driver than the human then it tries then it would apply it everywhere the problem is that's too hard the problem the in China the approach taken is let's take a simpler example use tremendous engineering and execution to build real working autonomous vehicles in constrained environments for customers who would pay money and then you collect the data and move to harder problems so from left to right here you see the Chinese companies are getting better and better starting from indoors with autonomous forklifts then going into autonomous Logistics transportation for airports which is outside then moving large cargo at ports with these giant trucks then to trucks that can drive on the highway but not yet in the city then in the city for buses these Robo buses are going on bus routes that might have 20 stops but it doesn't have to go anywhere then it goes to street cleaning which only drives at night when there are very few pedestrians out we want to minimize potential of hitting a human and there are many other examples and these companies are growing rapidly generating more Revenue than the typical American counterpart and I would predict in another two years we're going to see a number of Chinese autonomous vehicle companies that are widely deployed with government applications to transportations in cities and in highways on the bottom you see that various capabilities of AI in manufacturing I talked about how AI would take over 50 percent of routine blue color and white collar work the blue color work would be people who currently work in the factories for perhaps for visual inspection on the left side basically using human eyes and then for muscles and moving and then for hand eye coordination and for Dexter charity and more and more complex now ai is starting to conquer that and we as Innovation are looking at the types of workloads that AI can do and investing in companies that are gradually taking away the need for routine human labor in the factories China being the factory of the world has tremendous motivation to automate because that's the way to compete with countries that have a lower cost like Vietnam rather than lower the salary which is not practical anymore these are some examples of companies we invested in on the upper right you see these tiny little carts that coordinate with each other in a in in in basically a swarm they never run into each other because they know exactly each other's coordinates and you see that they can go so fast for very one important one very important reason humans not allowed right if human were to enter who knows what would happen because humans can make errors humans are unpredictable but these machines are fully connected like one swarm of bees or um Birds who never run into each other on the lower right hand side is something we're quite familiar with today a PCR tests here is a giant robot that can you can see on the lower lower right of the lower right how fast these arms are moving doing PCR tests right a human would be very cautious for the fear of contamination for the fear of making errors for the fear of squeezing too much or too little liquid but robots don't make mistakes they're fast they work 24x7 so this robot can do 120 000 PCR tests per day and that's how China can do the massive testing that we see happening today and then on the lower left is a non-ai breakthrough it's a materials breakthrough this is a robot whose fingers are not made with big steel construction but made with the soft materials it has these fat bubbly fingers you can see here it's grasping an egg yolk something that even human hands cannot do you can try it tonight if you want and then another important area related to AI is the semiconductor Revolution and the need for more compute people might say why do we need more compute my iPhone is fast enough my PC is fast enough what they are but there are new workloads like AI that is no longer finding today's compute power acceptable this graph shows um the most advanced AI applications for that year the x-axis is the year and then each Blue Dot on the graph is a an AI demonstration so all the things you've heard about alphago alpha alpha fold Alpha fold 2 gpt3 are all on this graph and what you can see is before 2010 the the AI applications were growing very slowly because AI was not known was not important AI was forced to conform and use whatever the fastest computers were available and computers between 1960 and 2010 were going growing at guess what 1.5 x per year that's known as Moore's Law right so whatever computer Intel or others make we AI people had to use them so we better not demand more compute because that's all we're dealt with but starting 2010 the AI became hot inventions came out data became plentiful they require more Computing AI became hot so that companies like Google and Facebook built 100 million dollar computers and super computers were made and gpus were invented so we saw a very fast trajectory from 2010 to today it's growing at seven times per year and imagine the power of exponentiation how much more compute is needed next year the year after the year after we're going to very definitely run out of compute very very soon the Y graph is the amount of compute and each little tick represents 10 10 times more compute so the computers we have are not going to get us where we need to be if we want AI to advance itself so what happens here is more explicitly the blue chart shows the CPU the green chart shows the GPU so think of it as an overall morsel if you will that with or without the GPU we more or less maintained enough progress based on conventional semiconductor chips a 50 increase per year which is tremendous already but what AI needs is the orange picture and that we don't have and projecting 2035 we're going to be off by six or seven uh six or seven orders of magnitude the computers that Nvidia will be able to make by year 2035 on this graph will be something like a billion times too slow than the AI scientists so what are we going to do are we going to be able to advance computers by a billion times magically well the answer is yes and that is by Quantum Computing Quantum Computing provides a complete different way to make computers rather than bits they have qubits so one of the magical things is that by adding a qubit you double the capability of a machine so a machine that has four thousand qubits by just adding three qubits to it four thousand and three qubits you've made a quantum computer that's eight times faster than the one you had last year thereby enough to to make the AI scientist happy so if we want AI to advance then we better push for a quantum computer if you doubt this think about who's the most powerful Semiconductor Company in the world right used to be Intel that was the PC days and then then it was Qualcomm that was the mobile days now it's clearly Nvidia these are the AI days I know that you don't touch and use AI or don't feel like you're touching use AI every day but this has already happened Nvidia became the most powerful semiconductor doctor company because the AI workload is creating value and changing the world but that's no longer going to be satisfactory because we need quantum so Quantum as I described will be the largest shift in Computing in the human history I would argue I have argued in our superpowers that AI would be the most important technology in the history of human race it's a little bit controversial but more and more people are accepting it but just as people are accepting it I need to correct myself I would say now when quantum computer works that will be the most important technology eclipsing Ai and revolutionizing AI the day that quantum computers start to work for any of you who have Bitcoins in the classical Wallet be careful they will be stolen because a quantum computer will break your bit computer past Bitcoin password okay but if you have a modern wallet don't worry you're safe but that's just one joke as it's true but it's a small example but Quantum will break all computer security every password everything you use will completely be broken um until computers start to use quantum encryption then they will be unbreakable even if you had a quantum computer to break it that's how amazing this technology is quantum will change the future of AI algorithms every AI algorithm we have today including deep learning is an approximation it's not an optimal strategy and having a quantum computer will enable a whole new class of AI algorithms that will completely leave deep learning behind in the dust that is why Forbes Magazine has said the future geopolitical competition in technology will be fought in the labs with brains on whoever figures out Quantum first which is why on these other charts you see us and China in investing vigorously into the space today us is ahead but China is developing also very rapidly the fourth area is life science I'm going to pick one specific part of life science which is Healthcare we all want to live healthier and longer and the the Breakthrough is about to have a set of breakthroughs are about to happen in healthcare Life Sciences first everything is being digitized including wearable Computing Radiology reports hospital records and genetic sequencing doctors will no longer be able to digest all of our data to give us the proper treatment I mean what doctor can breathe a gigabyte of genetic code is impossible but computers can do that and it can also treat people in a Precision manner just like when each of us logs into Tick Tock or Facebook or YouTube we see different content because those are designed for us when in the future when we get treatment we may each get a different treatment even if we have the same illness That's The Power of AI and it can expand on all kinds of ways of treating existing and new diseases secondly a lot of new technologies are being developed to make new Machinery new medical instruments new robotic surgery combining the power of mechanical robotic AI with the need for instruments and machines to become more powerful with these advancements and an example would be think about how blood tests were done when you were little right you were Pro you probably had a someone looking through a microscope counting white blood cells now a PCR takes care of it so we're at the point where these medical devices can be built that are smaller faster more accurate and cheaper so that will push the revolution one step further and a number of things that we invest in from the right the right hand side shows that AI drug Discovery is great it's because AI can read all the medical journals that humans can't and come up with drug candidates that are likely to work on the left hand side you see actually a company called Mega Robo that builds a giant robot on the top is the PCR one I told you about 120 000 PCR tests a day on the bottom is a more fancier version of that that automates the wet lab in the wet lab is where today the scientist or the technician truck runs experiments look depend on the results runs more experiments now the experiments can be done by a robot it's not about Job replacement and saving money it's about the scientists becoming more productive because 24x7 that obedient robot will run the program written by scientists yes if you're studying science you have to learn programming in order to tell your robot what experiment to do do and what you get you get eight to ten times more productivity because it's faster and there are other types of breakthroughs small molecule large molecule drug Discovery and so on lastly we talked about new materials and and and applying those to Green Tech is a big area we can make one molecule at a Time new materials that are cheaper it can be applied in agriculture and very importantly new energy using hydrogen and and sodium and other technologies will turn energy from a competition which used to measure who has the most endowment from natural resources like oil but now it's going to be who can make better solar panels and battery storage and China is way ahead in this area China has five of the top 10 battery companies in the world and eight of the top ten solar panels in the world on the right hand side you see a typical Chinese battery company that's going Beyond batteries you see a picture there that is making a battery that is the chassis of a car that will reduce the weight of the car and the price of a car by 30 to 40 percent and allow the car to be charged once and run a thousand kilometers imagine that future it might be a battery company through their new reference platform that challenges Automotive companies so in summary we're looking at in 20 years an opportunity for us to use these five revolutions that work with each other and have synergies with each other that will build a few whole new society in which there is enough for everyone and will have a chance to not only be liberated from routine work that machines can do for us but also live in an era of plenitudes where we will have enough resources and ability to create materials labor and energy that will feed everyone on Earth and hopefully eradicate poverty and hunger thank you [Applause] thank you Dr Lee for your insightful and really enriching presentation may I now call upon Miss ayushi Lahiri our student moderator for tonight yes ayushi the tambusian second year engineering student oh economics excuse me so without further Ado I will open up the Q a session I would just like to remind students that there are two microphones on either side of the room so when you um before you begin asking your question please introduce yourself all right and we would like to take one question at a time so if you do have a second question um you know wait for your turn again right so please don't be shy and come to the mice all right so without further ado ayushi thank you so much Dr Lee for a very insightful speech about I guess where the future of AI really lies and how this is really going to be benefiting Society um so I think I will start off the Q a session and then I will open it to the floor so um when you were speaking about how we could actually rely on AI so for example doctors doctors was the example that you'd used and how you'd have sort of a handy assistant to kind of just pull data from the internet that previously um doctors and scientists would have had to do manually wouldn't there come a point where your AI becomes so much more advanced than the supposed people that they're helping and if we reach that point what do we do about the quandary of like the need to be intellectually engaged for so many people that are off these professions to kind of have you know a job to do I see very good question um usually people ask about Job displacement which is related to the question but you went a step further about intellectual challenge so first on job displacement um I'll be clear that ai's first step let's say in the next 20 years it's going to do mostly routine work so if you're truly doing it intellectually challenging work then it probably would not be replaced by AI nor would AI threaten your desire and opportunity to be intellectually challenged on the other hand if you feel your intel if you're in the job and you feel intellectually challenged and yet AI is replacing your job I would challenge you by saying maybe your job is not intellectually challenging right all of us do many things that are intellectually challenging as an example a radiologist right they do many things that are very intellectually challenged they do research they do analysis and they interact with others they talk to patients that's all very important but the process of looking at an MRI scan and determining if it's cancer type A type B type c well I'm sorry that is not intellectually challenging we need to rethink what is intellectually challenging the fact that AI can do it means it's pretty routine because think about think about facial recognition right I mean I know it's a little bit controversial but facial recognition is something we think people are better at because we when we were babies we taught ourselves we wired our neurons to recognize our parents faces so we ought to be really good at that they ought to be innate because babies don't want to be separated from their parents but yet facial recognition systems can recognize three million faces with uncanny accuracy and we as humans can barely do a thousand right so think about how much AI is pulling way ahead of us in an example of something we used to think to be good for humans and maybe intellectually challenging but now ai is millions of times better than us but why do we think reading watching pictures of tumors and labeling them is an intellectually challenging task when recognizing faces are not it's actually not only not so intellectually challenging I would argue Radiologists are being taught a bunch of rote and inaccurate and erroneous sometimes rules that don't generalize and that is imperfect and AI can do a better job so and I would I'm not saying a radiologist's work is useless I'm saying the part of their job that is recognition of MRI is something that AI will do better it's done better for some domains it will do better for more so I think we need to look at this point and question with an open mind and in fact um all of us do spend the day doing some percent routine jobs perhaps I do 30 perhaps you do 40 uh perhaps a factory worker does 90 perhaps a radiologist a 60 so I would look at the flip side to say that percent of time for which we're doing routine work AI is being sent here to liberate us so we don't have to do them and our 30 60 70 of our time can be freed to pursue more intellectually challenging things and I would go as far as to say we might want to consider for a moment anything AI can do for the next 20 years is relatively routine and we should be doing we should be better we as humans should be doing things AI cannot do and and that's what makes our I mean if you think about people who who believe their worth is in their jobs right we don't want to give people jobs that I can do it better because they'll just be disappointed and and have their you know Ambitions destroyed we want people to Aspire we want people to happily give to AI to do things that AI is better at and to change their pursue to things that AI cannot do as a way to improve themselves and to challenge themselves and it's a gift for more intellectual Pursuits and yes a lot of your training may be thrown away a Radiologists spend Years Learning how to recognize tumors but at some point in the next 20 years not that soon by the way I'll be clear at some point it may be in the next 10 to 20 years that skill set will become useless just as you know switchboard operators who used to do this to connect telephone calls elevator operators who used to have to see how the level elevators right typists who used to have to type on typewriters and many many jobs that people spent years or decades learning and perfecting became useless when Technology Innovation does a better job and that to us to us needs to be a wake-up call that we shouldn't do that anymore we should move on we should not try to suppress or slow down that progress or or call it as a way to to hurt us in any way but we should move on because the there's more opportunities for you to to learn new skills that are more interesting and more challenging very interesting perspective because I think the the majority of perspective that were presented in the media is that the AI is coming to take away all of our jobs but you kind of speaking about how it's going to unlock new levels to human potential how we might be able to kind of scale past what we consider now to be intellectual tasks um kind of opens very many new doors for people to kind of explore and grow in those areas so thank you very much for that answer with that I would like to open the floor to any questions please remember to state your name thank you thank you Dr Leo for your insightful sharing I have a question here my name is Yong Yang I have a question here uh related to what Henry Kissinger the former U.S Secretary of State said in his book uh written co-author with your former colleague in Google he said AI processes are pretty fast and satisfying and there are some concerns about whether humans will lose the capacity for thought I mean say deep thinking or what he called Deep literacy being able to think in the long term and spend time in thinking things like that I'm sorry can you is is that a I didn't quite catch the quote can you so you say it again AI processes up really fast and satisfying you can really enjoy this kind of things but there are some concerns whether humans will lose the capacity for thought and conceptualizing and perhaps have and possess deep literacy I got it okay uh yeah actually I think this uh somewhat similar question um and and I I think um on the one hand I want to say that uh AI is really not having any Deep Thoughts AI is as magical as all the things I talk about AI is merely a glorified spreadsheet right it's got all this data in it you push a button magic happens and but magic happens because it's analyzed the data when you're having a conversation with an AI it might look real but it's really not thinking the way we think it's a different brain that collected all the things people ever said and typed in the whole universe and came up with a conceptual representation that tries to give you an answer that it thinks might be satisfactory it's actually not through uh deep thinking so um so I don't think that's that's uh the way AI works on the other hand I will acknowledge Dr Kissinger's point that AI is from a user perception more and more intelligent right AI can have a conversation synthesize text and read people's emotions and perhaps create the appearance of having an emotion when it doesn't have one and uh and that that will progress so I I don't think we yet know what will happen as people continue to develop more human capacity and and emotions and and AI that appears to be more human um my my personal view is that in the next 10 years let's say despite all the progress being made to masquerade as humans and appear to have deep thought and and empathy and and emotions masquerading as human is is is really hard uh it will if you make one small mistakes people will distrust you forever and people have also I believe an innate desire to interact with and reciprocate emotion with other people I think to the transference of that to a robot appears to be unlikely in the short term so in other words I think pretending to be a human with deep thoughts and affection and all that will be very hard to do in 10 years and even if it does a decent job AI will tend to have catastrophic mistakes in the because of the Black Box approach it takes so people will probably not accept it because it you know imagine you have a robotic girlfriend uh 10 years from now and it's wonderfully nice 99 of the time but one percent of the blue it says something horrible you'll probably not want to keep that robotic boy girlfriend right so that's what the AI will be like so at some point when will we get to 99.99999 that will be a long time because it gets harder and harder and then lastly even if you got to close to 100 people will have a preference interacting with other people for many things for uh for love for friendship for caretaking for feeling good about something and transference to the robot I think is an likely so all these things combined I think Dr Kissinger's fear are probably not within the next 20 years and um and and I think uh also we as humans who really should spend more time making AI that as a tool that solves problems and helps us and is symbiotic with us and is built around for human good rather than emulating all the things to make a fake human okay thank you thank you for your question Young hello I'm uh Jonah my question is are you've spoken before on the on restrictions in the Chinese information space and for the reasons that you've highlighted earlier it's likely that Chinese air companies will continue to rely on this Chinese information space even if they uh also harness data from outside so are you concerned that Chinese AI might be less culturally sensitive or more insular possibly even have racial or ethnic bias compared to other Ai and given the place of AI within China's broader geostrategy would that translate over to how China conducts itself as a whole okay let me answer the bias question I think all AI systems Chinese or American have inherent bias today that bias comes from two parts in a small part maybe by the programmer or the engineer who is not thoughtful or who is biased himself or herself and uh carelessly makes mistakes or intentionally makes mistakes that leads to bias outcome and I think that needs to be fixed by enhancing education for AI Engineers so that they're aware that with greater power comes greater responsibility that they need to be accountable just like the medical students have to swear the Hippocrates oath that life is sacred AI Engineers need some kind of an oath that they must not enhance create and grow bias and negative thinking in the society and here are the set of tools they have to use to limit that so that's a small part the big part of the bias comes from the data so so the example used in in the US is that a particular tech company used an AI algorithm based on its um interview results to recognize what kind of a job applicant would likely be successful in that company and in their training the AI model they used many more men than women so the AI algorithm basically learned among other things that the company wants more men than women then it continues to send more pass-through more men candidates and fewer women candidates and that's how bias happened in that particular case in another very famous case showed that facial recognition work well for um for Caucasians but poorly for African Americans and again that's due to a lack of training data size for that racial group so what is important is for us to make sure that data sets are reasonably balanced so that the the bias in data sets doesn't lead to biased AI systems and probably even worse example was a product called Microsoft Tay it was um chatbots that was allowed to talk to people on Twitter but pretty soon because it's trained on Twitter talk and you know how people talk on Twitter it started to uh you know insult people and make racial remarks and gender inappropriate remarks and Microsoft had to take down the system Microsoft didn't want to program a a biased system but because it learned on the data the Twitter data is really a problematic to learn from so we really need to conduct research on how to cleanse and select data and also to filter the output so that it doesn't exhibit those behaviors so those are the things that that that can be can be worked on um and and I think we need to have better tools for those who are in engineering you know that the compiler right you write a program a compiler runs through it it will give you a few warnings that why if you really run this program here are a few bad things that could happen uh run it at your own risk similarly when you compile an AI algorithm it could come back after reading your data but say hey your data is imbalanced you're likely to be discriminatory you're likely to have insufficient data and be inaccurate and so on so better tools better training awareness of the problem in the data coming up with research that better cleanses the data so these problems are minimized I think those are my my answers to the question okay thank you thank you for your question maybe we'll do some from the side okay um well good evening Dr Lee um I'm actually worried about receiving targeted scam messages in the future so I mean my question my question uh probably is how should we regulate AI or such a formation technology that you previously you mentioned from being used for malicious purposes so about government regulation of AI yes yeah I do believe regulation is clearly needed because of so much potential externalities that could happen but I would first say there are multiple Solutions uh we should not think of Regulation as the only Panacea and it's certainly not a Panacea I'll give you a couple of examples one way is to regulate AI by inventing new technologies think about when electricity was invented people got electrocuted there was no regulation preventing electricity going to the home but what solved the problem was the invention of a circuit breaker and when the internet was first connected to the PC you were all too young but take my word for it it brought a lot of viruses to the PCS but governments didn't stop connecting Internet to the PC they what solved the problem was antivirus software right so and we now know that these work when your email folders were crowded with Spanish M there was no regulation to to to to eliminate spam you can't do that but you know spam detectors put them in the stamp spam folder so these Technologies should be permitted and encouraged to be developed and applied because they often are the best solution to problems caused by technology secondly uh I I think as I mentioned earlier there should be better tools and better awareness by the engineers so that they have fewer of these bad examples that that would come out um and uh and and thirdly I think there can be uh third-party things that can eliminate bad things in AI rather how do you how do you regulate fake news are you going to find Facebook a million dollars for each piece of fake uh fake news it's very hard to regulate um and it's somewhat um not binary it's a little bit you know ambiguous whether something is a fake news or not however while the third party Watchdog can do is they could have a fake news index and then if a certain company lets that index go way up then people will discredit the company and not use their products so that function could be used it could be a third-party Watchdog there could be an ESG metric linked to responsible management by social network companies for fake news so that if they spread too much fake news some funds won't be allowed to invest in the company that's another way um and there are business model issues right we talk about well is Facebook really bad intentionally bad by showing us all these negative content or you know Tick Tock or YouTube uh what they're just trying just they're just trying to make money by using AI to show things that we think we will uh click and like but one of the big issues behind it is the advertising business model creates a unfortunate misalignment of interest between the giant company and us what we want is get good information learn and grow and be well liked and make friends these are the these are our interests but Facebook and YouTube and tick tock's interest is get more user minutes and make more money so that create and yet their business model being advertising causes them to show us more content that's addictive and but without regard for whether it's good for us or not so that misalignment is a problem some business models are less prone to that problem such as a subscription business model for example I don't hear anybody say Netflix is really showing us fake news and hurting our kids and our our minds because it's a subscription business model if if it does that to you you unsubscribe and that has a natural effect of putting Netflix out of business so they don't dare not to respect users interest and align with our interests so all of these things I think in combination need we need to do a better job then when these things can't solve the problem then we can take a a regulatory approach um and and Regulatory approach is is very difficult because you are you don't want to micromanage your company's business yet you want to protect the user and you also are dealing with multiple issues such as Anti-Trust and Monopoly maintenance and um and just not managing your AI well so so far probably the government that has come out with a larger set of restrictions and regulations that has achieved a fast effect is China I'm not arguing it's better or worse it's more of a fast execution of a set of things that it made the internet companies not do thereby creating a more competitive environment and um and and less opportunity to hurt the consumer so it's somewhat effective but it's also slowed down the industry right right the European approach is a an approach to focus on the user data and the sanctity of that and moving towards a way to um uh moving towards the way uh to protect data so that large companies can't get access to it so these are all the things you should consider I should add one more uh web 3 is argued as one possible way to get to to solve this problem by basically using blockchain to publish everything there is no data to take there is nothing to abuse things are exposed in the open of course that creates a whole new set of problems but it solves this one that we're currently talking about so basically my short answer summary is that there are a multiplicity of ways of dealing with externalities of AI and regulations one of them we should do it but we shouldn't think is really a Panacea thank you for your question um I think we'll do one more from this side uh try and keep your questions precise um we are running a little bit short on time please proceed it with your question and remember to State your names hi my name is Lee Rand and um first of all uh I just like to say I'm a huge fan of the book AI superpower and a lot of young people are impressed by your career achievements and your involvement Microsoft Google and synoventions so this this may be a bit more personal question but my recall may not be perfect but in AI superpower you said that in 2013 you were diagnosed with stage four cancer and it changed your frame of mind that money and fame when that important and that what was important is family and relationships so do you have career advice for like young people like us should we aspire to work as hard as you to achieve so much or should we value like family and relationships more yeah don't go to school and just go make lots of friends and get a girlfriend right no no no I'm just kidding um yeah I I think it's it's age dependent right that was my realization at age of 50 when I had cancer that I felt I'd spend my whole life optimizing my career I felt I was like an AI algorithm actually that I was trying to optimize my my my wealth and fame and um and and and and it creates a feedback loop and a very dangerous behavior and it's um obsessive and it takes away from all the other things in your life so what I realized when I had cancer was that when I thought I might only have a few hundred days to live I didn't want to spend any one of those days working I wanted to spend it with people I love and people who love me I want to do things I enjoy I want to satisfy my interest in hobbies and learning and curiosity I didn't want to work anymore so that my first 50 years were spent with an imbalance towards those things that were you know Fame and glory and wealth but not with uh real negligence towards the things that really important to me and and really it a book I read by Bonnie Ware was really resonated with me because she spent time with people in their death dying before they were dying and their largest regret was exactly the same as mine spending way too much time working spending way too much time obsessed on things that they don't really care about spending too much time caring about things the society wants them to care about and not doing the things they love not spending time with the people they love so this is something that is important but having said that I don't think everyone should remove their ambition and just um you know spend all the time with family and friends um I think there are smart ways to to spend there's time there's a way to work hard and work smart um similarly there's time to be smart about how to maximize your time with your friends and families right spend more quality time with your family not measure by the hours but by your focus and true sincerity and um and being thoughtful it's not how many hours you spend so young people should work hard but you shouldn't spend all of your time working you should balance between work and other things that you're interested in and as you get older you might get more temptation to work even harder because you want to accumulate wealth and success and achievement and fame but but keep in mind as you get older your time becomes more limited and it's actually time to focus more on work-life balance so my my advice to the young people think about the regret that I had when I was 50. what should you do now to prevent that regret when you're 50. not uh work less hard and you know spend all the time making friends and and being with your parents and and girlfriend but balance it too thank you for the question maybe we'll just do something yeah hi uh good evening Dr Lee thank you for your thought my name is Justin I'm here to computer science student um you mentioned in your talk that you know China is experiencing a rapid development in Ai and investment in Quantum Computing and you yourself are heavily invested in this area as well so my question to you is how can we ensure that ethics and ethical guidelines are developing as fast as so as to keep up with these developments in AI how to keep up with the advancements yeah ethical ethics and ethical guidelines your your computer science student and you want to know about how to make sure you're aware of these ethical issues yeah okay no that's a great question I I think it's precisely the problems we see today with these large internet platform companies abusing AI is precisely because they're Engineers didn't ask you a question that their management said quarterly result get more user minutes convert more revenue and then that led to a cyclical negative behavior so um so there are a number of Publications and websites that you you could and should look at for example uh I think there are a number of organizations I chair World economic Forum Council on AI and then there are a number of other AI for good organizations and there are some excellent books and videos that have been created uh one one of my favorite people is Tristan Harris and he he made the Netflix video social dilemma and I think that's a really profound and very simple and anybody can understand and I would say also if you're a computer science student why why not pursue ethical AI as your direction right I don't mean you know going to philanthropy and make no money I'm I mean you could work on that compiler too I talked about what if you worked on the tool what if you went to a company what if you went to Google right or Facebook that worked on their tool tensorflow or or Pi torch on a function that looks at all the AI models being generated and issue warning messages for the AI That's creating bias or other problems right that kind of a career career path you can be very successful accomplish a great deal and at the same time be a positive Force for good in this area thank you thank you for the question we'll do another one from the side hi yeah my name is Andre year one psychology student I just wanted to ask a little bit about I think you mentioned just now about the reduction of the percentage or at least replacing the percentage of routine work right so that humans can be unlocked to do what is more intellectually challenging to us right I'm just wondering how can we convince people that are skeptical about AI to reconsider their relationship between themselves and using or how do you call it uh knowing that AI is in the background of most of the things that we do okay yeah I think there are maybe two types of people in this category first those who are who are who are aware but they're fighting back and they're in denial or they're just anti-ai and then there's the kind that are maybe not so knowledgeable uh but but large percentage of their workload is actually quite replaceable so what can we do with both these two groups of people uh with the first group people since they're knowledgeable they've done some homework I would try to persuade them that AI is a tool it's a spreadsheet on steroids it's something we can control and manage it's something we can use and make ourselves better I would make analogies like when Photoshop came out would artists feel oh my God that's taking away time from my my Canvas OR my camera my photography but if they viewed it as a tool right today artists and photographers who don't use Photoshop all right the great disadvantage in fact the ones who use Photoshop earliest got the greatest hit you know hit the ground running right and similarly the writers who used typewriters and didn't move to word processing and word and the computer I fell behind because much of their time was wasted um you know dealing with the mechanical tool that can't do error correction so being the first to embrace technology gives you an advantage and uh fighting back the tell them to read about the ludice Revolution right the luddites you know they they were against technology they burned things in the end what happened they lost their jobs so they should spend the limited time they have on learning the skills that will obviously help their career rather than be denial or in the naysayer against it against the technology that is inevitable they're definitely going to lose right if you're you know a horse carriage Carriage driver in the in the early 1900s and you can do all the Lodi things you want to do against automobiles but it is a wave that cannot be stopped so anyway I would try these things and if it doesn't work it doesn't work the second group of people which I worry much more about is the people who do currently routine work and a significant percentage of a workload will be replaced by Ai and furthermore their the routine work they do is not a very skilled type of routine work and they're just unaware that this tidal wave is coming and when it comes several things will happen they'll lose their job suddenly they'll try to grasp at any straw any job that they can find also likely to be in less skilled work and then guess what in a two or three years their job will be replaced again they will go down on this down downward spiral very worried about about that some people talk about uh Universal basic income let's just give everybody money well that doesn't solve the problem the assumption that people who didn't know this tidal wave was coming lost their job and now we just give them money they'll find the right direction that's not likely most likely they'll need a lot of assistance and training so I would say that societies and countries should put aside sizable training budgets and programs to help people identify jobs that are going to stick around and retrain them in the right directions and for example it's pretty obvious right if you're an auto mechanic that job's going to be gone the automobile of the future is going to be you know electric and autonomous vehicle is going to be like a phone on the giant with a giant you know windows and and wheels uh the skills for for the current auto mechanic is doomed uh with or without AI right and job of a plumber is reasonably safe because that's a tough one it has to do with you know when is if I poke a wall and and and the the the the the the the the lady who called me to do the work got mad what do I do and and when do I feel confident to you know break this up and where's the is it like a detective work that one's hard to replace and and there are many many such things that it's not hard to draw a picture of the likelihood and time for displacement and give warning to those whose jobs are likely to be displaced and help train them in jobs that are likely to be sustainable that's something that each of you can do if you know someone in such a profession and the government should do and universities should do and also vocational schools should stop training more more out in the automotive mechanical repair right they should put they should have more plumber schools and even more robotic repair schools right and things like that so so there are many many things we can do for that group of people and there's a giant group of people too so the impact of society is huge if we're seeing 20 30 percent uh people who lose their jobs for the next you know 10 in let's say in 10 years uh helping their uh getting settled is is a huge Challenge and that Education and Training process really needs to start as soon as possible and and lastly I think I forgot to mention I happen to believe one of the non-ai related jobs that will become more in number and and in importance and in Social stature are the human to human service jobs so Health Care Service Elderly Care and also other types of service jobs that require human interaction including people in sales people in doing concierge or tour guide and many new professions that require the human interaction I believe machines will not replace the need in fact it will accentuate and increase the need for human human interaction so we should guide more people in that direction as well thank you thank you for the question good evening Dr Yi I'm Ming tiang I'm a computer science and math student I'm interested in startups and Tech so I picked up your book a couple years ago and I just want to say that I found it quite insightful I'll try and keep my question short but um in your book and doing your talk you mentioned some of the benefits of the US where they have a lot of smart engineers and they attract a lot of the best talent but obviously China is beating them in terms of having an amount of data tougher engineers and sorry more data tougher entrepreneurs and a surplus of Engineers obviously I'm now convinced that China's leading the world and I'm applying for my Visa as I speak I'm just kidding but I understand that you definitely moved from the US to China in 1998 I believe and um you've obviously achieved a lot of success with the growth of China but increasingly China is looking inwards and does not seem like I'm gonna get the same good chance if I do choose to move to China today so I'm just curious as a student who's interested in Ai and startups and I'm sure for some other people in the audience like what opportunities do you think there are for people that are not based in the US not based in China and what sort of advice would you give for students in our position thank you okay uh I think there are actually many opportunities actually you're quite fortunate to be in Singapore right if you think about what big companies have lots of data and are beginning to make good progress on AI well you guys have Tick Tock you have grab you have C these are all great companies that equivalent companies to them five years ago were the ones that carry China forward right it's if you think about all the great startups that were founded many of the people came from the the baidu's alibaba's and tencents if those were the Giants of China you now have a few Giants in Singapore and and furthermore all three of these companies are Global which is incredibly important because if you're in a small country and all you do is work in a company that addresses your domestic local market then you will never have big data you will also not have a lot of users your company also won't be worth that much so Singapore hasn't made a giant lead forward by having multiple companies that have international presence with data from those countries that can accelerate their AI development so my prediction is five years from now if I were to come here as a VC looking for companies to invest engineers in Tick Tock and grab and see are the ones I would be talking to because they learned a great deal about using big data in AI companies with a lot of data and they're the ones who might be doing the AI Healthcare startup or the AI energy startup that's exactly what happened in China you can actually replicate that in a reasonable scale here because you're now training some of the people who are starting the Chinese startups in these areas so I would say assuming you want to stay in Singapore go work for one of these three companies or one of the other companies that have sizable data and X excellent AI programs and learn a lot there if you want to be an entrepreneur then you can learn a lot and go leave and do a startup if you just want a good place to grow your career you can probably stay there for a long time because these companies are likely going to have good growth in the future question we'll do one from this oh hi Dr Lee my name is Andre I'm a YouTube computer science student and I'm also Italian so you are definitely a figure that I looked up to so my question will be more alerted to your work right now sign of ventures in China I would like to ask what is your approach in investing in China right now and how do you validate the technology in Taiwan in inside in China right now and also I have another thing is do you believe there is um I'm so sorry we can't have two questions just one I'm really sorry I do okay okay uh yeah uh we are now uh investing 100 in deep Tech in China um partly because I think uh the success of Chinese AI companies gives us confidence and gives the entrepreneurs confidence they can also be successful in other areas and partly because looking at this geopolitical technology competition it it's problematic in that it's inefficient for the whole world to have two ecosystems on the other hand it's giving both countries U.S and China a lot of incentives to go push and accelerate its domestic technology ecosystem in the exact area as I talked about Quantum semiconductors AI robotics Life Sciences new energy Etc so um so I think the entrepreneurs in both U.S and China in these deep Tech areas may get a booster and an accelerator because their respective government due to the competition will want to subsidize and help and push forward and accelerate so these are kind of the exciting areas to be if if 10 years ago in China mobile was the most exciting thing if seven years ago o2o was the most exciting thing if three years ago AI was the most exciting thing I think now we've got a number of exciting things so it's a different group of entrepreneurs who will tend to be deeper in technology many with phds in technology and that they want a VC to invest in them help them find the business Direction and and create Great Value so you know one of my most fun moments are finding really super smart people to work with and to find complex Technologies and make them work for for the for everybody so I feel that the opportunity to transition in my company from investing in any company now to invest in just the comp tech company is a way to use and magnify my own personal skills and interests in a global picture that may be a little bit gloomy sometimes but a silver lining is that deep Tech may get accelerated and fourth Industrial Revolution dream may come true in both U.S and China and since my company and my work is in China this is the hemisphere and segment in which I'll happily invest more years and hopefully create many more Great Tech unicorns thank you thank you for your question um okay uh maybe all questions all right all right um so do you want to Rapid Fire your questions then would you want to do that I can take these four questions but let's try to be be short okay doctor a year one computer engineering student and also a huge fan of your work in AI um so my question is on the tech rivalry between us and China so recently there's been escalations in the tech conflict with U.S restricting the export of high-end Nvidia gpus and also eov chips technology to China so I'm just curious as to your thoughts on how this rivalry will play out between us and China and whether it will end in peace or War I mean okay technology War I understand uh I think we're definitely headed for greater competition more intense competition and um uh I I think the when us tries to say we're going to make some products or Technologies unavailable for China it's a double-edged sword right clearly in the near term it will create a an inconvenience if not a serious um negative impact but in the long term it's basically an open invitation for China to enter exactly the space in which U.S won't give the Technologies so the the the balls in China's court right can China develop technologies that U.S will not give China any more and if it can China's huge Market will will allow China to build companies that it couldn't be built before because imagine can China five years ago create an Nvidia I would have said no but if U.S is going to say we're going to make this layer and available to China if you want it you get better build it yourself and then maybe next year another layer then effectively you're outlining a roadmap of saying well I'm going to try to suffocate your access to my Technologies I dare you to create your own so China is at the crossroads if it successfully creates what it used to depend on U.S for these Tech it will accelerate its path to a superpower and maybe challenge us and compete in other parts of the world but of course if China doesn't succeed then it will suffer as was intended by these policies so these are quite a double-edged sword the as far as general gpus you can already see China now has a dozen companies developing gpus and um so so I I think um the the obviously the jury is still out I can't make a prediction which way it will turn out but I think it's an interesting uh Gamble that the US has chosen to take and I think the balls in China's court and we'll see how it plays out in the places in which it shows GEOS chooses to play that game thank you for your question thank you let's take all three questions at one go oh um hello I'm Jose um I want I would like to ask do you think the Chinese government will be using ai's China AI China's AI superiority to further their political interests both domestically and internationally and like if if so I do have like any concerns about where it could be hated I'm sorry use it where um like do you think the Chinese government would be will be using like the China's AI superiority to further their political interests and if you have if you think so right like you have any concerns about where this could be hated all right thank you good evening I'm ramu thank you very much for your talk so far Dr Lee I would like to ask if you believe that they'll be Equitable in universal access to AI technology besides from our two main AI superpowers Us in China would first say the middle income countries be able to utilize these Technologies to adjust their challenges thank you thank you very much question thank you totally my question is a very simple one how and why did you write a book and what was some challenges of writing a book which is best selling all right very different questions uh I I I don't know anything about government policies so I can't answer the first question maybe you can invite another speaker from the Chinese Embassy in the future um on the uh the the second question uh I don't think U.S and China are dominating the technology access yes the Chinese U.S companies are leading companies in technology areas um but the access to these Technologies are widely available interestingly the large companies while they are sometimes very tough in their competition with each other they realize that in AI if one chooses to be closed another company might compete to make it open so Google's tensorflow was pretty close so Facebook said we're going to make an open pie torch right and then open ai's large language model was pretty close and Facebook said well we're going to make an open source language model so now these large companies are competing on who's more open so it's a great opportunity for University students and smaller companies because there's more open source stuff out there of course there's already GitHub and the open source code area so I think you you're in good shape relatively speaking when I was entering computer science however was very close so now you enter a much more open in terms of Technology access on the on the book the primary reason was I felt AI is so important I want more people to understand it and have access to it my first book AI superpowers while it had a number of air key points in it one of the interesting feedback was people felt I made AI relatively accessible so my question was how can I make it even one further step even more accessible and and then one day we had an idea that what if we turned it into fiction so AI 2041 which I didn't cover today is actually 75 fictional so I outline all these Technologies and the roadmap for the next 20 years and I gave it to a science fiction writer an excellent writer Stanley Chan and he turned it into 10 stories and he and I iterated on the stories to make sure they're faithful to my technology predictions and that they're interesting to read and after each book I would write the analysis explain how the technology Works what are the upsides downside and how to deal with the downside so it was the most fun I've had writing a book because I love movies I love science fiction even though I didn't write the stories here it was fun being a part of it and I'm hopeful there would be a TV or a movie coming out of this that would I would really enjoy going to Hollywood to to do that so and we're looking into that so it was it was great fun and it's great fun working with a creative mind I also respect back science fiction writers greatly I both respect them greatly and also sometimes wish they would be different I wish they would be different because sometimes they describe they like to describe Ai and technology in such dystopian context that causes people to be negative I'm a little frustrated by that but I'm very inspired by them because almost everything you see today in the world that's an exciting technology that we're we've now taken for granted probably you can trace back to the roots of signs some science fiction that because those guys are more creative than us us being the technologists right the writers are more creative they could think further and put scenarios that both serves to inspire us as technologists to work towards that Vision so we can hang our hat somewhere so we're not working like when I was working on speech recognition I wasn't working on it in that by at that time no speech recognition works I didn't work on the problem for the sake of publishing papers right I did it because I liked Star Trek I liked how Captain Kirk you know talked to to the Starship Enterprise and how they the agent was always very intelligent and interactive and fun and I dreamed that one day I could make that kind of capability happen and and it did happen and and many other people talked to me about how they traced the many other AI scientists tell me they were inspired by some science fiction which gave them the context and scenario and inspiration and the belief that if they took their research to fruition they would impact the world in the way that a particular science fiction did so that's why I have tremendous respect and very thankful very inspired but sometimes a little frustrated by science fiction and it was fun to be now a pseudo-science fiction author even though my co-author wrote all of it great thank you so much um thank you so much for the great conversation thank you ayushi thank you Dr Lee for fielding this wonderful questions um May I now invite Mrs Lima Rector of tambusu college for her closing remarks I think this can go on forever right um thank you so much Dr Lee for a wonderful mind-blowing session thank you for taking us painstakingly through the evolution of AI you know and as well as pointing us to all the different possibilities I mean which are really mind-boggling thank you for reinstating our motivation so that we do not have to be afraid of what AI potentially can do but in terms of how we can be a participant in all the different developments and possibilities that AI will bring to our lives so thank you very much Dr Lee for your time let's show our appreciation [Applause] [Applause] thank you thank you [Applause] thank you Mr Slim we have come to the end of the session I hope you enjoyed the session um you may take your lead hi everyone [Music]

Channel: TembusuCollege

Views: 2,249

Keywords: Tembusu, Tembusu College, Tembusu Conversations

Id: nqxtKjUZUAg

Length: 90min 40sec (5440 seconds)

Published: Thu Oct 27 2022