[Highlights of JD Capital · Future Forum] When Will “Algorithm” Rule the World
Speaker: Yuval Harari, the famous Israeli writer, author of Sapiens: A Brief History of Humankind and Homo Deus: a Brief History of Tomorrow.
The following is the script of his keynote speech (with edition) on the Third JD Capital · Future Forum held on May 25, 2017.
Today, I want to talk about the next revolution in human history.
Looking back in time, we have experienced many revolutions in technology, politics and economics. Despite of all revolutions, one thing remains changed, the human itself. We are still what we were thousands of years ago, just a kind of animal. We still have the same bodies, brains and spirits with our ancestors in ancient China or in the Stone Age.
Nowadays, though capable of controlling the outer world, we know little of how to control our inner world. We know how to stop rivers by building dams, but not how to stop our bodies from aging. You know how to kill the buzzing and sleep-disturbing mosquitoes, but not those annoying, rolling thoughts that may deprive you of a good sleep.
The next revolution will change this. It will not only alter tools, politics, or economics, but also human beings – it will bring changes to our bodies, brains and spirits. We can control our inner world. The main products in the 21st century will not be cars, weapons or food, but bodies, brains and spirits. We will learn how to design and produce them, which means we probably will be the last generation of homo sapiens. Within one century or two, the rulers of earth will be a completely different group, far different than what we are from the Neanderthals.
If we succeed in the design and production of bodies, brains and spirits, this revolution will be unprecedentedly great not only in history but also in biology. For the past four billion years, the basic rules of this game of life have not undergone any fundamental transformation. For these years, the organic compounds, amoebas, dinosaurs, tomatoes or human sapiens, have all evolved through natural selection. Yet this might be replaced by intelligent design. So, life will not be limited to organic formations. Cybernetic human or completely inorganic forms might be manufactured. An era of inorganic life through intelligent design might be waiting for us, the organic lives shaped by natural selection over the past four billion years.
How will this revolution take place? Its basics is the integration of biotechnology and information technology. On the one hand, we’ll gradually deepen our understanding of biology, genetics, bodies and brains; on the other, we’ll manufacture computers with better performances and computing abilities. Combined together, these two technologies will enable us to discover all mysteries of organism. If we have enough computing ability and data of human physical features, we’ll be able to create an algorithm which understands human even better than human itself.
This will change our economy completely. Now we believe that our customers are always right; in economics, the desires and choices of customers are paramount. However, before long we will see that algorism is right. It will shape our desires and make choices for us.
It starts from such simple issues as choosing, purchasing and reading a book. In the past, you did it based on feelings, tastes or friends’ recommendations. But now, people rely more and more on Amazon. When I enter the virtue store of Amazon, an algorism comes up and tells me: “I know you! I know what books you liked. Someone sharing similar senses with you likes this new book.”
This is just a beginning. Currently, Amazon’s algorism is only based on crude data which is easily misleading. A case in point is that you buy a best-seller after hearing someone praise it but only to end up with reading no more than 50 pages. But since you do not want to look silly, you tell everyone you really like it and even rate five stars on the Amazon website. Nevertheless, this will lead Amazon to recommend you and other millions of customers these “annoying” books. Now such equipment as Kindle will soon tackle the problem with the data collected when you’re reading. Your Kindle will see where you read fast, where slow, where you stop and at which sentence you give up this book forever. You may tell the whole world that the book is a wonderful masterpiece. But your Kindle knows you never finish the 50th page of it.
More surprises are awaiting. Engineers are developing software that can detect human emotions through the movements of eyes and facial muscles. After equipped with such software and video camera, Kindle will know what makes you laugh, sad or angry. Then, after connected to the biosensors on or within your body, your Kindle will know how the sentences you read influence your heart rate, blood pressure and brain activities. Soon, when you read books, books read you as well. What’s more, you might forget most of the contents you read shortly after, but computer program remembers. After you finish reading Tolstoy’s War and Peace, you might have forgotten most of it. Yet, Amazon will know who you are, what kind of personality you have, and how to trigger your emotions. Such data helps Amazon to select books for you precisely, and even to make the most important decision in your life for you, such as who to marry.
Algorism not only helps us make decisions, but also performs better than human beings in many tasks, or even replaces us. Take transportation as an instance. The idea of pilotless automobile once was a scientific fantasy, but now it is going to come true. Computers will drive cars faster and better than human beings. Every year, around 1.3 million people die of car accidents, twice the number of victims of wars, guilt and terrorism. Most car accidents are aroused by manual mistakes, such as drunk driving, texting during driving, or even falling asleep. Pilotless automobiles don’t have such problems, and therefore would reduce accidents by a large extent.
What’s more important, pilotless automobile can connect with each other and form unitary networks. Currently, cars are independent from one another. When two cars approach the crossroad from different directions, they cannot interact with each other. What they can rely upon are the limited signals which, if mistaken, will lead them both to the crossroad and then crash. If all cars are connected to a network, they will no longer be independent vehicles relying solely on traffic signals. Instead, they will know each other’s tracks and thus refrain from crashing accidents.
Of course, this also means that millions of truck drivers, bus drivers, taxi drivers, and even traffic policemen will lose their jobs. Once all cars are automatized, there will be no need for traffic police anymore as computers never break laws unless you order them to.
Atomization can replace not only taxi drivers, cotton spinners, but also teachers, lawyers and doctors. The major task for most doctors is to diagnose and offer optimal treatment. If the doctor tells me I have symptoms of a fever or diarrhea, I might suffer from bromatoxism. Yet other diseases like stomachache, cholera, dysentery, malaria, cancer or unknown diseases might also have the same symptoms. And since the limitation of my health insurance, he only would spend a few minutes to diagnose, answer only a few questions, or do a quick health examination. Afterwards, he would refer to my medical history and all diseases human beings have now. It would be a pity that even the most diligent doctor could not remember all my past diseases or checks. Also, no doctor is able to get familiar with every disease and medicine, nor is he able to read every essay in medical journals. What’s more, doctors sometimes get tired, hungry or even ill, which will influence his judgment. This will inevitably leads to misdiagnose or improper treatment.
Today, IBM, Microsoft, Apple and other companies all want to create AI doctors much more capable than human doctors. First, AI is able to read information of all known diseases and medicines from its database. Then, it also updates these data every day, adding into the base the most updated researches and also health statistics collected from every clinic and hospital of the world.
Second, AI doctors can rapidly access my genome and daily medical history, and relevant genomes and medical history of my parents, siblings, cousins, neighbors and friends. They will know immediately whether I have visited a tropical country, whether I have alimentary infection, whether my family members have intestinal cancer or whether everyone in the town has similar symptoms this morning.
Third, AI doctors never get tired, hungry or ill, so they can always work for me. I would be able to sit in the sofa comfortably and ask hundreds of questions, expressing my feelings. In fact, AI doctors would know my health condition even before I feel painful. It could, through biosensors, monitor my blood pressure, heart rate, blood sugar and brain activities 24 hours a day. Therefore, they could diagnose influenza or cancer at early stage and eliminate them easily and low-costly.
Currently, many hardships in technology are hindering this process of replacement. Hard though they may be, they will no longer be any problems once solved. Comparatively, the training of human doctors is complicated, expensive and time-consuming. After the training, another ten years are still needed to cultivate a doctor. If you want to train another doctor, then you should start all over again. So, if the technological problems are solved, plenty of AI doctors would be able to serve in every part of the world every day. From a long-term perspective, the cost might be as much as 100 billion dollars; even so, it is much cheaper than training human doctors.
Some believe that though the algorithm can defeat doctors and pharmacists in terms of professionalism, it can never replace their humanistic care. If CT shows you have cancer, would you like to receive the information from an indifferent machine or a doctor that cares for your emotions? Fine. Then how about from caring machine whose commands are tailored to your emotions and personality? Emotions are biochemical phenomena enabling AI doctors to check your emotions as accurate as your tumor.
Another possible revolution in health care might happen to the major target of health care itself. In the past, health care aims at curing illnesses. In the future, the aim might be changed into updating and perfecting healthy people. This harbors positive potentials, but also unprecedented dangers. Human might develop into different species, a great danger which must will be. With the development of biotechnology, it might be possible to prolong life and enhance living capacity. Yet, this new wonder might be expensive, not free for billions of human beings. As a result, human society in the 21st century might the most unequal one in human history. For the first time, economic inequality would transfer into biological inequality. For the first time, the upper class would not only be richer than others, but be more talented and live longer as well.
In the 20th century, equality is the core value of human society. Most years witnessed struggles for equality by working classes, women and national minorities. While in the 21st century, the upper class may pursue immortality at the cost of equality. In the past, the only differences between the rich and the poor were in social status and economic levels, not biological ones. In the future, however, significant biological differences might surface.
What seriously threatens human equality is that governments and elites may lose the initiative to invest in public health, education and welfare in 21st century. In the last century, both democratic and autocratic countries invested a large quantity of capital into health care, education and welfare as they needed millions of healthy labors to work in factories and millions of soldiers to serve the army. This was the same case with the Nazi Germany, who even built hospitals, schools and sewage systems for the poor.
But in the 21st century, the public might lose their military and economic values. This already happened for soldiers. While millions of normal soldiers were recruited to form in the last century, nowadays, only a limited number of elite soldiers and an increasing number of unmanned planes, robots, network worms and other automatic weapons can make a strong military force. Now, most people are useless in military. And this will happen soon in civilian economy.
As computers are performing better than human beings in driving trucks, diagnosing and conducting many other tasks, billions of people might be eliminated out of the employment market. Many new jobs might appear, but this does not guarantee a solution. Human beings have two basic skills, in physics and in cognition. If they are outperformed by computers, similar situations will happen again in new jobs just as in the traditional ones.
In the past, when automation altered employment market, people shifted from one low-skilled job to another. In 1920, a farm worker lost his job due to agricultural mechanization could find a new job in a tractor factory. In 1980, a factory worker lost his job could start his new career as a cashier in the supermarket. These were feasible because such shifts, from farms to factories, and from factories to supermarket, were from one low-skilled job to another, which required only limited training. But in 2040, a cashier or a cotton spinner, without necessary skills, once unemployed due to AI, would be hardly able to restart his career as a software engineer designing the virtual world. This issue would be particularly serious in developing countries as they rely too much on low-skill jobs for economic development. Are we training Bengalese young people to be software engineers? If not, how are they going to make a living in 2040?
As a result, human beings in future might break into a brand-new super-elite class and a huge useless class: billions of people would lose their jobs and no longer be able to get one. We have no such an economic mode to cope with this situation. How to manage this useless class might become a most difficult economic and political issue in the 21st century.
We are confronted with another huge danger. When learning about how to understand and design bodies, brains and minds, we would not only get the power to control the inner world besides the outer. However, they are different stories to control and to understand. For the past several thousand years, human beings have gradually achieved the ability to control the outer world. Yet, due to our ignorance of the complexity of bio system, we use our power unwisely which results into a nearly broken bio system. Similar thing might also happen to also inner world. We would probably get the ability to control and shape our bodies and brains. However, we are far from understanding our minds. So, the biggest danger for us may be that we, while trying to control and shape our inner world, would destroy it, just as what we have done to the outer bio system. After 50 or 100 years, we might then find the ruins of our own minds as they would not be ready to cope with those changes.
Lastly, I want to emphasize that what I describe are not predictions, but possibilities. No one know what the world really looks like in 50 or 100 years. Technology is never definite. We can use the same technological breakthroughs to create different societies and environments. For instance, in 20th century, human beings utilized techniques brought about by Industrial Revolution—train, electricity, radio, and telephone, but three different kinds of societies were created: Communist society, Fascist one and Capitalist one.
In the 21st century, the rise of AI and biotechnology will definitely change the world, without a single definite result. We can utilize the technologies to create a different society. However, how to utilize appropriately is the most important issue we face today. If you do not like some of the scenes I depicted just now, you can still work for some changes.
Thank you for your attention.