Yuval Harari - The Challenges of The 21st Century

so hello everyone in the 20th century in the past century humankind has really managed to do the impossible and reign in famine plague and war today for the first time in history starvation kills fewer people than obesity plagues kill fewer people than old age and violence kills fewer people than accidents and it's good to remember these amazing achievements as we look forward towards the new impossible challenges of the 21st century and the 21st century will be full of new and even more difficult challenges than we have ever encountered before ranging from climate change to the rise of disruptive technologies like artificial intelligence in this talk I want to focus on what is perhaps the most complicated challenge of all which is the challenge to our humanity in the coming decades the twin revolutions of biotechnology and information technology or really the merger of biotech and Infotech into a single overwhelming scientific tsunami this might very well undermine our conception of the of humanity and will shake the humanist foundations of modern civilization because when we come to confront any big challenge whether it's nuclear war or climate change or AI we always need some ethical basis to stand on and for generations our most solid ethical basis has been humanism but in the 21st century humanism itself might become obsolete now what exactly is humanism to put it very simply humanism is the idea of the belief that human feelings of the ultimate source of authority when we confront any big question of dilemma in our personal lives or in our collective life as a society humanism expects the feelings and free choices of human being to provide us with an answer humanism tells us to listen to ourselves to follow our heart to be true to ourselves and since no one can understand my feelings and free choices better than me no one should have absolute authority over me over my life this is the basic idea of humanism now this sounds a bit fuzzy and abstract and complicated so I'll give a few examples of what humanism means in practice because I think when we talk about such big questions especially ethical questions clarity is of utmost importance if you don't have clarity it becomes very difficult to really understand what we are talking about so what is humanism in practice let's look at several different fields let's look first at politics what is humanist politics humanist politics believes that the voter knows best and governments should serve the voter if you encounter any big political question you should ask the voters what they feel about it the feelings of the voters of the highest political authority and mind you that in referendums and elections people are not really being asked what do you think they are asked what do you feel if elections were about human rationality there was no reason to give everybody equal voting rights because different people have different rational faculties rational capabilities but in feelings supposedly everybody are equal and that's why for example when Britain needed to decide whether to leave the European Union they didn't go the Queen of England to make the decision they didn't ask the Archbishop of Canterbury to make the decision they didn't even ask the great professors of Oxford and Cambridge to make the decision no they went to each and every British citizen and asked him or her how do you feel about it and when the citizens said they felt like leaving the EU there was no higher authority that could tell them your feelings are wrong let's humanist politics the voter knows best now what is humanist economics humanistic anomic says that the customer is always right and businesses serve the customer how do you know if a product is good or bad it depends on the customers it's very very simple a good product is a product customers buy a bad product is a product customers don't buy it's as simple as that if you make a product and you're convinced it's the best thing in the world but nobody buys it it means it's a bad product you can't counter the customers and say your feelings your choices are wrong at least if you believe in humanist economics now how does this manifest itself in art and aesthetics what is humanist aesthetics just as humanist economics believed that the customer is always right humanists aesthetics believe that beauty is in the eye of the beholder throughout history there are many attempts to define to find some objective definitions for out and for beauty but then came humanism and says no there are no objective definitions there are only subjective experiences and subjective feelings in 1917 a century ago Marcel Duchamp took an ordinary mass-produce urinal which I think you can see there they're declared it a great work of art placed a tenant in an art exhibition and ever since then there is this raging debate is it art is it beautiful who decides what is art and if you're a humanist you will eventually conclude that well this depends on how people feel out is anything that people believe is art and beauty is anything that people find beautiful if somebody feels that this is a beautiful work of art and is willing to pay millions of dollars to have it so who is there in the universe that can tell this person you're wrong it's not art it's not beautiful so this is human is aesthetics now what is humanist ethics humanist ethics in essence believes that if it feels good do it again throughout history there are many attempts to define some absolute an objective morality which is independent of human beliefs and feelings and experiences based perhaps instead on divine revelation that in the Middle Ages homosexuality was considered a terrible sin because the Bible said so because the church said so because the Pope said so then came humanism and said that we don't care about what the church says or the Pope says of the Bible says we care about human feelings this is the ultimate groundwork for morality if two men love each other and their love doesn't harm anybody why should anybody think that this is bad that this is evil of course sometimes out dilemmas even in humanist ethics what happens if the same thing makes me feel good and makes somebody else feel bad let's say I steal your car I feel very good about it but you feel very bad about it so in that case we don't apply to some higher absolute morality we have to weigh the feelings one against the other this is how debates moral debate in a humanist society happen we weigh the different feelings one against the other and we usually conclude that theft or murder are wrong not because some book said so but because they hurt people they make people feel miserable that's why they are wrong and finally in a brief survey of what humanism is what is humanist education humanist education teaches people to think for themselves and in past generations in past eras like in the Middle Ages the main aim of education was to teach people what the wise books of the wise men of the past thought what the Bible said what Aristotle said because this was the main source of authority but when humanism rose the purpose of Education change humanism says that authority comes from within yourself not from outside and therefore the main aim of Education at least in humanist societies changed and if you go to a teacher from kindergarten to university and you ask what are you trying to teach your students so the teacher would say I try to teach history or chemistry or physics but above all I try to teach them to think for themselves because this is the ultimate source of authority so this is humanism and it has dominated our world for quite some time now but in the early 21st century humanism is facing an enormous challenge not from dictators of demagogues but above all from the laboratory's humanism against says that authority comes from our feelings which reflect our free will we nobody besides us can really understand but science now tells us with greater and greater falls in authority that this is all a myth this simply isn't true there is no free will it's a myth feelings don't reflect free will feelings of biochemical algorithms and given enough data and enough computing power an external system can understand me much better than I understand myself the big idea of our era whether we like it or not the big idea of our era is that organisms are algorithms and algorithms can hack organisms including Homo sapiens it's just another organism again organisms are algorithms and algorithms can hack organisms now this is even more complicated in the neumann ISM so again let us try to explain and give a few example what does what are the scientists mean when they say this well first of all North Lee will so the Natural Sciences biology chemistry in physics they understand the world better and better far from perfect but still better and better and as far as they know in the universe there are only two kinds of processes there are deterministic processes and there are random processes and randomness isn't freedom that's very very clear there are we are not familiar with anything in the universe which can be described as a free process free will as far as we understand today is an empty concept that doesn't describe anything in reality or in nature humans certainly have desires they have a will but they are not free to choose the desires to choose the world and their feelings therefore don't reflect any kind of free choice of free will and certainly their feelings according to the life sciences are not some spiritual quality that God gave only Homo sapiens in order to appreciate beauty and in order to make moral judgments no all animals all mammals all birds have feelings and these feelings evolved by natural selection as biochemical algorithms for making decisions diam they are based not on three intuitions but actually on calculating probabilities the big debate about the heart vs the brain emotions versus logic actually there is no debate emotions feelings sensations they are all actually also calculation which happens so fast that we just don't notice the calculation which is happening in a split second below our level of awareness let's look at a concrete example a baboon stands in the African savannah and sees a tree with bananas on it but not far from the tree there is also a lion and now the baboon needs to make a decision a decision of the kind that every animal needs to make every day of its life do I risk my life for the bananas or not and the survival of the baboon depends on this decision and in order to make this decision the baboon really needs to calculate probabilities what is the probability that if I don't eat these bananas I will die from hunger versus the probability that if I try to get these bananas the lion will eat me to survive the baboon needs to make a good calculation of probabilities now how does in order to to to make the proper calculation the baboon needs a of information information about the bananas how many bananas two or eight of they're big or small green or ripe information about the lion how far is the lion how big is the lion is the lion asleep or awake does he look hungry or saturated and also the baboon needs a lot of information about himself how hungry I am how fast I can run and so forth and you need to take all this information together collect it analyze it weigh the probabilities and reach a decision how does a baboon does it the baboon doesn't take out a pen and a calculator a piece of paper and start calculating know the entire body of the baboon and especially the sensory organs and the nervous system and the brain this is the calculator within a split second the baboon takes in smells and sounds and sights and sensations from within the body and then billions of neurons in the brain process the information in a split second the calculation is made and the answer will come not as a number but as a feeling an emotion fear or courage or perhaps in the sizes indecisiveness confusion fear and courage are the way that the calculation is manifested and they are not spiritual insights they are a biochemical calculation and this is true of how baboons make decisions about bananas this is true of how British citizens make decisions about brexit and this is true about how German citizens make decisions about immigration what's happening there is the very quick biochemical algorithm making a calculation there is no free will or spiritual insight involved now the humanism was wrong to think that feelings reflect free will until two day until the early 21st century it still made very good practical sense to believe in humanism even though there was nothing magical about our feelings they were still the best method in the universe to make decisions and no outside system could understand what's happening within me and how and why I make these decisions nobody had the biological knowledge necessary and nobody had the computing power necessary to make sense of what is really happening within me and why I feel the way they feel even if say the Stasi followed you around 24 hours a day 365 days a year all the time looking and watching and eavesdropping on the conversations still the Stasi did not have in the 1960s or 1980s the necessary biological knowledge of what's happening inside the brain of the human being and the Stasi did not have the computing power necessary to make sense even of the information that it did manage to gather over you and therefore it couldn't really understand you but now the merger of info tech and biotech is changing the situation advances in biology and especially in brain science are giving us some of us at least the necessary biological understanding and at the same time advances in computer science especially in machine learning and AI are giving us or some of us the necessary computing power and when you put the two together when Infotech and biotech merge what you get is the ability to hack human beings there is a lot of talk of hacking these days about hacking computers and smart phones and email accounts and bank accounts but we are really live you're entering the era of hacking human beings and once you can hack human beings then Authority is likely to shift from human feelings which are no longer this black box that nobody understands the authority might shift from human feelings to computer algorithms and humanism Zanden dilections and the free market and so all this will make no more sense we see it happening already today beginning to happen first in the field of economics we are living behind the era of the customer is always right and entering the new economic era of the algorithm is always right because the algorithm can predict and manipulate the feelings of the customer let's start with a very simple example from the book industry how do I choose which book I the consumer the customer of the book industry how do I choose which book to buy so in the past in the humanist era I rely primarily on my own feelings and literary taste but now I increasingly rely on an algorithm that at least allegedly knows me better than I know myself as I enter the virtual Amazon bookstore the first thing that happens is an algorithm pops up and tells me I know you I know you based on your previous likes and dislikes and what you bought and then and so forth and based on everything I know about you and millions of other readers I recommend this or that new book to you now of course it's still very very primitive and the Amazon algorithm makes a lot of mistakes because it doesn't really know me so well it doesn't have enough data but it constantly gathers more and more data and is improving and its gathers data in new ways if you read a book on an electronic device like Amazon Kindle the Kindle can read you while you are reading it for the first time in history books read people more than people read books with far greater attentiveness anyway is you read a book on Kindle Kindle can know which pages you read fast which pages you read flow and which page and when you stop reading and this gives the algorithm and much better idea of what you like and what you don't like but it is still very primitive the next stage which is technically feasible even today you can connect Kindle to face recognition software and then the algorithm will know what makes you laugh what makes you cry what makes you bored what makes you angry the ultimate step is to connect Kindle to biometric sensors on or inside your body and then the algorithm will know the exact emotional impact of every sentence you read you read a sentence and the algorithm knows what happens to your heartbeat to your blood pressure to adrenaline level to your brain activity by the time you finish the book let's say you read Tolstoy boring piece very long book by the time you finish Warren piece you forgot most of it but the algorithm will never forget anything by the time you finish this thousand page book the algorithm knows exactly who you are what is your personality type and how to press your emotional buttons and using this kind of information it cannot just choose books for you with far greater accuracy than you can it can also tell you what to study and where to live and whom to marry and people are likely it's an empirical question if you learn that it gives you good answers you increasingly rely on it until you lose the ability to make decisions yourself we see it already happening with like navigating space that people have learned that it's good to listen to the smartphone to Google Maps if they need to get from here to the bus station to the train station just listen to Google and after they do it for a while they lose the ability to navigate space by themselves and even more dramatically one of the most important abilities of human beings is to look for information to look for answers to questions that bother them but more and more many people know of just one way to look for answers to question just as Google within a very short time of say 15 20 years this crucial ability has been out sourced from the human mind to the algorithm and there is no reason to think that if the algorithm is good enough it will not happen in the same way with the ability to choose what to study or whom to date or whom to marry now an algorithm can hack humans they will be able not just to decide things for them but also to replace them and in the coming decades the twin revolutions in info tech and biotech are likely to disrupt the job market and might lead to the creation of an enormous new useless clause a class of people who are useless not from the viewpoint of their mother or children nobody's ever useless from the viewpoint of his loved ones but useless from the viewpoint of the economic system there is nothing they can do better than an algorithm than a computer than a robot and it should be emphasized this is not the result just of the rise of artificial intelligence by itself artificial intelligence is not able to to do something so truth of dramatic it's the merger of Infotech of AI with biotech because for many jobs in order to perform the job well you need to decipher human feelings and human emotions computers and robots will never be able to replace human doctors and teachers and lawyers and even drivers unless they are able at least to some extent to identify human emotions correctly and some people have this may be wishful thinking that this is something computers will never be able to do they can make calculations but they can't understand emotions they can't understand feelings though they will never replace human doctors or teachers now this makes sense if you believe that feelings are some mysterious supernatural phenomena that God gave only Homo sapiens and that works in some metaphysical way but if you accept that feelings are just a biochemical pattern which in the end is a process of calculation there is absolutely no reason to be so sure that a I will not surpass human beings even in emotional intelligence it will not have feelings of its own it will not have consciousness but it will be able to know that this person is now fearful of this person is now angry with far greater accuracy than any human doctor or teacher or lawyer now just as the authority of the algorithms might come to replace the customer and the worker in the economic sphere they might also come to replace the authority of the voter in the political sphere democracy is ultimately based as we said earlier not on human rationality but rather on human feelings in elections voters are not really asked what do you think they are asked how do you feel and if algorithms can hack human feelings and manipulate and predict human feelings then democracy is likely to become an emotional puppet show politicians or at least some politicians are a bit like musicians and the instrument they play on is the human emotional system a politician gives a speech and there is a wave of fear all over the country a politician tweets and there is an explosion of hatred and this fear in hatred this is the fuel of a lot of political systems what will happen when these musicians have a much more sophisticated instrument to play on what might happen is the rise of digital dictatorships in the 20th century democracy defeated dictatorship because democracy was better at processing data and making decisions we tend to think about the conflict between democracy and dictatorship as a conflict between different ethical systems but it was also a conflict between different methods for processing data democracy works as a distributed data data processing system democracy distributes the information and the power to make decisions between many institutions and organizations and individuals dictatorship on the other hand is a mass method for processing data in a centralized way democra dictatorship concentrates all the information and all the power in one place now given 20th century technology it was simply inefficient to try and concentrate too much information and too much power in one place it didn't work well nobody was able to process the information fast enough and make good decisions and this is one of the main reasons why for example the Western bloc defeated the communist bloc in the Cold War there is a story that in the last days of communism in the light eighteen 1980s a Soviet official came to London to try and understand in the capital of Margaret Thatcher how does a free society in a free market actually function and the British hosts they took this Soviet official on a tour of London to visit the banks in the stock exchange and to meet all kinds of economics professors and and other luminaries but after a few hours the Soviet official exclaimed wait a minute there is something I don't understand back in Moscow our best minds are working on the problem of how to provide bread to Moscow and nevertheless in almost every bakery and grocery store there is such a long queue for bread here in London a city of millions we've been going back and forth across the city for hours and I haven't seen a single bread queue so please cancel all my other visits and appointments and just take me to meet the person who is in charge of providing bread to London I must understand his secret and the British house they scratched their hands and they looked embarrassed and they said there is nobody nobody is in charge of providing bread to London that's the secret of a free society and a free market nobody is in charge you just allow the information to flow freely between all the different parts and to allow individuals and organizations to make their own decisions but it is not a law of nature that under all circumstances centralized data processing always is less efficient than distributed data processing this was the case in the late 20th century but now the revolution in Infotech especially machine learning and AI may swing the in the opposite direction machine learning and AI might make it possible to process enormous amount of information centrally and actually the more information you gather in one database the better the process the better the data processing and then the main handicap of authoritarian regimes in the 20th century their attempt to concentrate all information in one place it could become the main advantage in the 21st century and this could result in the rise of an completely new kind of regime very different from the dictatorships of the 20th century a new kind of digital dictatorship finally when algorithms can hack organisms they could also start creating and redesigning new kinds of organisms new kinds of living entities it's quite likely that the main products of the 21st century economy will not be in textiles and vehicles and weapons they will be bodies and brains and minds we are learning how to produce and engineer and manufacture them we are learning how to design and create new organic beings by spinning up natural selection in a way we are learning how to create cyborgs which are beings that combine organic with inorganic parts and finally we are even learning how to create completely inorganic beings and if this indeed happens this will be not just the greatest revolution in history since history began about a hundred thousand years ago it will be the greatest revolution in biology since the very beginning of life four billion years ago for four billion years nothing fundamental changed in the basic rules of the game of life many things happened the dinosaurs appeared the dinosaurs disappeared the mammals appeared all kinds of things but the basic rules didn't change all beings for four billion years whether amoebas or dinosaurs or tomatoes or Homo sapiens were subject to the laws of natural selection and were subject to the laws of organic biochemistry made of organic stuff now in the 21st century natural selection might be replaced by intelligent design as the basic driving force of the evolution of life not the intelligent design of some God above the clouds but our intelligent design and even more so the intelligent design of our clouds the Google cloud the Microsoft cloud they will be the main driving force of evolution and at the same time after four billion years of being stuck in the small puddle of organic biochemistry life might break out into the vastness of the inorganic reel so who in decide what to do with these godlike powers of creations of creation will the voters decide will the customers decide when we just listen to ourselves and follow our heart but how do you follow your heart when your heart is constantly being monitored and manipulated by an algorithm that's the big question of our time thank you [Applause] so thank you you all for this grand history and this grand image of how we have to look forward for some challenges within the 21st century and as far as we have running out of time at the very beginning of our conference I just would like to invite one person to raise a question and I asked Sheila jasanoff who unfortunately have to leave us immediately after her talk that she could have the opportunity Sheila jasanoff from Harvard so Sheila it's fee thank you so much for giving me this opportunity and it's obviously a daunting task to follow what you've just said there with any questions whatsoever I'm tempted to ask about 50 different questions beginning with the fact that you're useless class slide was highly gendered and it made me think that perhaps there's hope because there weren't too many women in the useless class but but visuals visuals aside you know ethics which is what we're celebrating today has taken root in spaces of doubt ambivalence disagreement not knowing ways forward your talk is was extremely the opposite of that I mean that is it's your talk was I'm trying to bring up something misbehaving sure well I may have to look for it again I don't know why this is happening I may not be able to give you the exact quote but your point about demagogues playing Publix and like musical instruments it brought to my mind a passage which some people in this audience may be familiar with it's out of Hamlet I wanted to quote the exact words but my iPhone is not bringing up the exact quote at this moment but you recall he has a conversation with Guildenstern in which he hands Guildenstern a flute and says play this flute for me and Guildenstern keep saying my lord I can't I mean you know I don't have the skill I can't play on the keys etcetera etcetera and Hamlet says look you do you think that I am easier to be played with than a flute which you do not have the technicality the technical procedures to deal with you are algorithmic future that you're outlining not only has mastered the flute business mastered the human and I just wonder where in your you know space of deterministic futurism against which there is a lot of evidence as well you have room for the kind of challenge that Hamlet was posing to Guildenstern who was presuming to know algorithm like what would make this very ambivalent Prince tick so what would you say how would you rewrite that passage and Shakespeare I wish I could read it for you because I would love you to produce the piece of counter Shakespeare for us in the moment well it's a very very good and timely question I'll try to answer a few aspects of it because it's a very complex issue first of all in order for this kind of scenario to be realized the algorithm doesn't need to play the flute or the of the person perfectly it just needs to do it better than the average human in order for us to trust Google Maps to navigate the city Google Maps doesn't need to be perfect it just needs to be better than the average human in order for self-driving cars to replace human drivers they just need to drive better than the average human they don't need to drive perfectly that's an impossible that's a - nobody can do that of course but better than humans is not such an impossible mission today for example human drivers kill about 1.25 million people every year mainly due to human error so to improve on that is not an impossible mission and similarly for algorithms to play a greater and greater role in making very important economic and political decisions they don't need to be perfect they just need to be better than the average human which is again not as impossible as sometimes we tend to imagine and I don't say that this is a good future or a good development we have to be extremely careful about the terrible dangers it involves but to be really aware of the dangers we need to be more humble about our own abilities as human beings compared to the computers compared to the algorithms if we fortify ourselves inside this kind of imaginary fortress that oh we have things the algorithm will never be able to understand we have free will we have the spirit we have the soul they will never understand that we are safe then it will be extremely easy to do pass in the ludus the easiest people to manipulate for example with fake news is people who trust too much in their own freewheelin and say oh I'm not making this decision because of anything it is just my free decision there are the easiest people to manipulate because they can't even conceive how easy it is - once you get to know the weaknesses once you get to know their hatreds or their fears it's all easy to manipulate them unless of course they have a much better understanding of themselves and of the weaknesses and of their imperfections I think that when we come to confront this big danger of the algorithmic future a more realistic and humble appreciation of human ignorance of human stupidity of human weakness is is key I my main fear is not in the end from artificial intelligence it's from natural stupidity I think it's a far greater danger because this is the opening that artificial intelligence can go through now I definitely don't think the I don't believe in technological determinism that okay AI is coming and there is nothing we can do there is just a single outcome this can result in some creepy dystopian future no no technology is ever deterministic we always we don't have free will but we do have will we do have choice we do have some power left and we know from history that the same technologies can be used in very different ways if you look at that East Germany in West Germany they had access to exactly the same technology they had electricity they had radio they had cars they had trains they just chose to do different things with them it's the same ways AI and biotechnology they will change the world but they can change the world in different ways and it's still up to us to try influence the direction it is taking and I think that as I said a more humble and realistic appreciation of human abilities and human limit limits will make it more more not certain but will make it more hopeful that we can make the right choices thank you

Channel: The Artificial Intelligence Channel

Views: 199,176

Rating: 4.7779155 out of 5

Keywords: singularity, ai, artificial intelligence, deep learning, machine learning, deepmind, robots, robotics, self-driving cars, driverless cars, Yuval Harari, Sapiens

Id: FSloTpkHYYI

Length: 46min 12sec (2772 seconds)

Published: Wed Aug 08 2018