Visions for the future | Kai-Fu Lee and Chen Qiufan

[Music] foreign Tech I'm Simone Ross the tech curator here how will artificial intelligence change our world last year AI expert and former head of Google China kaifu Lee and Stanley Chen one of China's top science fiction talents co-wrote a book to explore just that it's called AI 2041 10 Visions for the future each chapter is an imaginative creative story accompanied by an analysis of the tech and an explanation of why such science fiction might actually be plausible stories of separated twins who have their own AI companions or lovers who have never met over 20 years of pandemic living with each story offering a unique framework to think about what our future could look like I interviewed kaifu and Stanley at a TED membership event in 2021 foreign [Music] thank you so much for joining us for today's event featuring technologist and AI expert Dr Kai Fu Lee and science fiction writer Stanley Chen they are co-authors of the new book AI 2041 10 Visions for our future which will be released here in the US on September 14th kaifu Stanley thank you both so much for joining us kaifu I'm going to start with you um why write this book together well it is my belief that AI is the most important technology for mankind in the history of mankind and it's important for everyone to really understand it its implications and challenges but just reading technology descriptions such as the ones I have written as a technologist and investor is just not reachable to everyone and I think it's so important that if there is a way to make the storytelling really interesting and engaging and even into entertaining that more people can access it and can kind of see where the future holds so with that idea in mind I talked to Stanley about co-authoring the book with him writing the stories and that's what hopefully will draw a lot more people to it wonderful and assembly um I believe part of the process was you wrote the story first and then kaifu would write the accompanying sort of explanation did that constrain your imagination or your creativity in any way knowing that there had to be an explanation of how the technology would actually work I'll say the process is not exactly like what you imagined so I think both kaifu and I we built up the road map together we'll discuss in depth like how AI technology will evolve over the next 20 years and which points of Technology we should package and put into different stories and then I come up with some directions attached to each each story and then we will discuss and back and forth and then I and get out to write a story and then kaifu will write a tech analysis accordingly so I think from the very beginning from scratch we are collaborating very close and I don't think it will restrict my imagination but the other way around it trigger and inspire a lot of my imagination based on the realistic facts of AI Technologies I used to work by myself so I can do whatever I love but now it seems like the conversation and interaction really brings some chemistry into the story it's really like dancing right so so a Stanley and kaifu dancing duet is is what we have here great so I'd like to understand a little bit um kaifu from you you know we've been talking about the potential of AI for such a long time now you know decades even and um what's different now well you know I know that more than everyone else because I did my PhD in 1988 and then I worked for Apple Microsoft Google my CEO always asked me when will AI be real and that's really just happened in the last five years if we look at how much progress has been made to date this is not going to 2041 this is just today think about alphago alphago zero Alpha fold gpt3 autonomous his Vehicles the use of AI in all internet companies the use of AI and financial applications and all the robots that's working and all the RPA Technologies that's um displacing White Collar work is really been blossoming in the last five years and and I think that is just the beginning for for much more is to come in the next 20. so a future is finally here um we have some questions from the audience so this one is from estefania do you think there's a limit to AI since it's a human creation or will it reach a point of self-development yeah I'll take that first so um AI actually is not completely programmed it's programmed to have a goal it's programmed to have an architecture but then it learns by itself on more data and the more data that's fed as long as you have enough processor the better it gets and that is what has enabled the huge progress in the last five years it's not just smart people inventing new algorithms to solve all these problems but it's people framing it in a certain way and then letting the machine learn for itself so the self-organization self-learning and the fact that it gets better with every 10 times increase in data it gets better that is the amazing and sometimes scary thing about AI now about ai's ability to self-program I don't doubt that will happen to some extent but so far AI still requires a certain level of direction from a programmer and then it will learn under the architecture set out by the programmer so in the context of the book AI 2041 I think you will see AI that is highly intelligent uh very interactive gets better than you think on a lot of capabilities and it keeps growing but it's not growing as many people would assume in previous science fiction that it becomes a superset of humans but rather it grows in many directions on many things we cannot do but humans still maintain you know our soul our creativity our ability to love and and that that is I think something that I don't see replacing by AI in the next 20 years so so we will sort of maintain our humanity and and we won't be etched out by the AI Stanley do you have um some thoughts on that so I think it all depends on the three key factors like computation power like the algorithm and also I think the data so as kaifu mentioned like right now I couldn't see that like the singularity Point were coming over the next 20 years but as a science fiction writer I'm pretty looking forward to that day even maybe it will somehow like conquer the human beings okay conquering human beings that's that sounds a little scary um okay another question from from the audience this one from Medan um I believe this is um directed to kaifu in a recent interview with Peter diamandis you were talking about longevity and 150 years lifespan how do you reconcile this prediction based on Science and Tech advancement with the reality of ever increasing prevalence of chronic diseases at a decreasing age onset okay I I don't think I quite made 150 a year projection Peter may have I'm a strong believer in that AI will collect will use a lot of data much more than medical doctors and scientists have ever been able to collect including our genetic sequencing our MRI and scans and the full blood tests with all the markers and with all that data coming in from a large number of people it will be able to compare any individual to a prototypical healthy person and measure the Gap and figure out ways to improve whether it is sleep or reducing stress or eating better or nutrients or medicine or exercising that is something that is not so far off I'm actually personally experimenting with the company we invested in I have been collecting all of my data extensively and for the past year and I've been measuring my my blood for now against people who are at my age and five years younger 10 years younger five years older Etc and it's given me very sensible advice now with a human doctor interpreting the AI output of course and telling me okay you're really not exercising enough we should not think of Aging as the inevitable irreversible reality we can't live forever our body has a you know certain expiration date but I think there are many ways to make ourselves healthier it's not it's also people say oh I don't want to live forever I'll be miserable I'll be sick the whole idea is not longevity for longevity's sake but it's how to get you to be younger more energetic and more healthier and and that's correlated with living longer so I believe this this will happen and um I I don't know how long we can extend our age so there's no reason in the next 20 years we can't be in the mid high 80s in another 20 years maybe in the 90s even up to 100 that's definitely uh imaginable based on what I have experienced and and the Gap that I think remains to make us healthy in a Precision medicine kind of way that just hasn't we just haven't collected all the data we're in a health-wise we're in the Yahoo days right Yahoo days of the internet that's where we're at we haven't yet uh discovered Google Facebook and all the rest so we're just collecting the data and I think the best is yet to come I love that we're in the Yahoo days um another question from the audience so while we're talking about improving things um this one from Monica what's the impact of AI for education so I I believe uh if you look at all the industries education is one that has changed the least over the last hundred years right the way we communicate by Zoom entertain ourselves with you know iPhones and VR and work working remotely is completely changed the scenes totally different education is still one which is mostly a teacher giving a lecture to all giving a test to all giving homeworks to all it's um uh it's basically one size fit all kind of approach and that's understandable historically because one could not afford you know a teacher for every student but I think going forward I think we want a teacher for every student because every student has different weaknesses different preference is different Hobbies some would be would score better if we cast the problems as a basketball or as dancing or something that excite them and also if you can imagine a an AI teacher companion that is watching over each student helping with them and making learning almost a gamification a fun experiment so a lot of the content can be taught by this personalized targeted AI companion teacher right we think about targeting as a bad thing because you know Facebook targets us with things that makes us addicted or extreme in views but if it's targeting us to help us learn then this aligned with our goals I think that is a huge power of AI that can make a one AI teacher per student kind of a future and and the human teachers don't disappear at all you still need human teachers to help with things like encouragement personal connection create understanding values and learning creativity and encouraging critical thinking and teaching about communication and teamwork and EQ and all that then it becomes a much much better education much more targeted and also personalized maybe Stanley can talk about the the story in which he talked about AI education yes Stanley go ahead yeah I highly recommend everyone who had interest in how AI could change the education industry should read a story twin sparrow is talking about two often boys who were adopted by two different family which both use AI companion teacher in different way so it's not only about customized algorithm for each student but also I mean now we have this kind of online class like every day on Zoom but it feels so isolated so people might feel lonely when you spending a lot of time on a platform but actually you couldn't feel each other but in in our imagination with MPL natural language processing and XR technology like AR via Mr so actually you can have this kind of realistic of presence of each other and it's not isolating each other but to reconnect each other so I hopefully when you finish the story you'll feel the future of education is so hopeful and so warmth and full of empathy and passionate of of our like next Generation future yeah right another question from the audience um this one from Fabi I'm going to paraphrase it a little bit um so what is the impact of AI on poorer countries how do how do we we ensure that that this is somewhat evenly distributed kaifu can go first sure I think in the short term there are already efforts in place to make sure that big countries don't just come in and take all the apps and make them used locally and take all the data away and extract money without giving anything back so I think having some short-term laws and regulations to ensure that if you're a Citizens data is being used to monetize some big country's app that you're properly compensated for it and then that money can go back in the economy because really really data is incredibly valuable and that near-term measure needs to be done in the in the longer term I believe each country needs to develop its specialty and and as I described in the book it's really not just about AI sometimes it's about cultures that have special characteristics you know some culture really value craftsmanship others really value giving back others value service others value family and I think globally speaking AI will end up creating a lot of value a lot of economic value and it's not just AI but also uh the advances in um synthetic biology and energy and new materials so things will become cheaper and better so that in the future more and more people won't have to do routine jobs for a living they can't do what they they like and and they can contribute positive energy and also the ideas can help globally so I think you know some countries might want to focus on tourism and and the exciting things about its civilization and the storytelling so I think a lot of those will become possible I think a service industry will become huge so you know today people are using products in you know branded invented in the U.S manufactured in China which is why these two countries are a little bit a lot more wealthier right now but but I think in the future what will be of the greatest value is not just the products because products will become commoditized with all the things we describe in the book and what we'll value the most is service and human touch service and things that have strong cultural elements so I think those are possible directions as well and of course all that said every country should make sure people who are strong in learning about technology and algorithms and AI should get to do so so that the country does not fall behind because this is a incredibly important technology for you know look at how much industrial revolution has helped us become a leader in the globally and and I think AI is another potential equalizer or unequalizer so countries should really value maybe not AI for everyone but for those who are interested and and skilled you've said that the AI should be sort of a great equalizer but also potentially not so Tom is asking why we should trust AI scientists to develop these powerful Technologies and I'd like to add on to that that you know why should we trust our governments and and Regulators um to help get us to a more equal society as they claim to want to do well I'll take a shot and then Stanley can can add two of that um well we should first make sure everyone's educated so we can all become participants Watchdogs and raise our opinion if the programmers and scientists and corporations and governments aren't doing their job hopefully the book describes clearly that many of these problems of AI are extra actually externalities they're not intentionally creating harm for the individuals but they're done as a side effect so I think the key thing then is um you know AI scientists are bad people don't trust them but it's rather make sure they're educated to understand with greater power comes greater responsibility so what must they do when they program an algorithm to make sure it doesn't have bias or unfairness or treat you know women or minorities unfairly Etc uh so and also better tools should be developed so even if a scientist is negligent it catches it up front um and also you know with a large internet companies uh some of what they do is perhaps uh just out of greed others are maybe um not well thought out and some of it is just unintentional so there needs to be Watchdogs I don't think we can abdicate our responsibility and say hey government you fix the internet companies you know break them up and punish them and make them pay big fines that doesn't solve the root of the problem the root of the problem is that AI is so powerful with this algorithm so powerful that large companies cannot resist using it to make money um and unintentionally and sometimes intense at our expense so I think the key issue is how to align the large company's interest with ours okay so it sounds like you think that we collectively also have somewhat of a responsibility um and role in driving where this sort of AI future can take us so Stanley I have a question a question for you this one is also from the audience from Hannah what AI advancements that already exist do you think are Stranger Than Fiction yeah I think there are a few things actually make me think like science fiction writers is kind of lack of imagination in reality so for example like the fake there's something I don't think in previous science fiction had a lot of description on the thing like how people using technology to transfer their their face their voice their movement to pretend one another meanwhile it's not for good but it's doing bad things and how we can fight against those misbehaviors so another thing is like right now I think AI is more capable than human being on recognizing Michael expression even detecting all these small influx instructions in your voice in your tongue and I can foresee what your mental status is and what kind of reaction and response you am I gonna make so I think fundamentally we will learn a lot from AI because it's a mirror is like human beings but it can reveal so many hidden structure and hidden Knowledge from this huge amount of data which a human being as as a collective group couldn't really interpretate uh deeply enough but AI can so moving on to something where AI probably does have a big impact um another audience question this time from Jen in the next 20 years will some aspects of Industry be fully replaced by Ai and will AI cause unemployment is so easy to be replaced by machines in another a story called the job savior and it actually set in California San Francisco is talking about how people should be taken care of when there's Mass scale of uh structuralized unemployment happen and along with the development of AI technology like invaded into different Industries so our thought is like besides basic income that's one solution but I think work is much more than making money it's about dignity it's about self-actualization it's about how you finding your position in the family and among others in the society so we have this kind of idea we should use and leverage AI Technologies to do the job allocation if we can live up to 150 so it's no way that you only do one or two or three jobs right so I think AI should totally work to help people to retrain and reallocate in a new job and to help them to find their own value in this uh highly uncertainty Dynamic change of job change this question comes from a bunch of people um what scares you most about Ai and the future I think the most scary thing about AI is it it can be an amplifier to blow up all the negativities in humanity so maybe you you've heard from the news like there's a troubles on Twitter for Microsoft and it learned all the cursing and bias and discrimination during the interaction and conversation with the trolls so I think there's something very scares me because during this kind of feedback loop it might trigger even more like darker side of human beings right now we are have this kind of extreminist on the internet on the social medias but whether we understand if they're talking with the childbirth or not so maybe it's some terrorists were using this way to trigger some extreme emotions and protesting virtually so I think that something scares me in the future what excites you most about AI so Kai through you first and then Stanley well I've already talked a bit about um getting AI aligned with our interests in some sense it's addressing Stanley's concern that AI would Target each person and make us think things that are against our interests but but let's suppose for the moment that we can somehow get at least a set of people to program AI in a way that aligns itself with our future interests I think that's something that's incredibly exciting um you know Tristan Harris talks about it none of us know how to do it it's perhaps a new app ecosystem maybe it's a new device we're all willing to pay money for things that enrich us things that make us happier and more knowledgeable so if only we can figure out how to incentivize people to build that app and get monetized for it and and you gotta align the economic interest to make that happen that's probably the uh one very important thing that excites me the other thing I would mention is that everything is double-edged sword we talk about Job displacement routine work being done by AI but I would ask that at the end of the day right in 20 years let's say all the routine jobs are or 90 of the routine jobs are displaced by AI does that put us in a better or worse place right assume we help people transition from routine jobs to better jobs and we educate our prostate prosperity in a responsible constructive way so that people are pursuing their interest and their talents and not repeating what's routine then we will have done something that is phenomenal for mankind namely that we will have liberated Humanity from ever having to do routine work so that we can do things that we were really put on this Earth to do so that's probably the number one thing that excites me but but the road to get there is treacherous and dangerous and full of potholes Stanley what excites you I think if my uh totally changed the way we do the scientific research for example now we're using Alpha full to predict the protein structure and we're using computer vision to dig out the the connectum of human brain so I think there's a lot of things ahead of us a way for AI to help human being to understand more about the fundamental mechanism of ourselves and the world so I think that's that is a paradigm shift in the future so with the help of AI I think we're gonna fix a lot of unsolvable problem including climate changes so I think that's very not positive future ahead of us but I think we should should hold on to that positive imagination because we try to create the future we should start from Imagine One well I think positivity is a great way and a great place to end so we go a is going to hopefully liberate Humanity um from drudgery and free us to do the things that we will put on the earth to do and it is going to hopefully help solve things like climate change for us I think that pretty much wraps up our time for today so thank you both so much for joining us um we really appreciate it and thank you kaifu and Stan bye-bye bye