Yuval Harari: It's up to us to choose whether new technologies are used to build heaven or hell

 　　“After I wrote A Brief History of Humanity in the summer of 2011, I felt I would never tell that story again. I have a special fondness for that book and am grateful for its success, but I feel I have finished telling the story of the human species today. Human 2.0 is still unfolding, but I think the story in the future is best left to others.

　　"But, after the 2016 US presidential election, I think I'm afraid I have to go back to the beginning, Revisiting the unique ability of our species to construct structures of order and domination through imagination.

　　...

　　"The great challenge we face today is to create a new imaginary global order that is not based on nation-states or capitalist markets. Is there a way?

　　" That's the story I want to tell in this book. "

　　It wasn't me, Yuval Noah Harari, who wrote these words, but a powerful artificial intelligence system that was instructed to imitate my writing style. This artificial intelligence system is called GPT-3, and it is a San Francisco family. Made by "OpenAI", a laboratory specializing in machine learning. GPT-3 was instructed to write a new preface for the 10th anniversary of the publication of "A Brief History of Humanity", so it collected the books and articles I have written, The interviews I've done, and the billions of sentences I've found online, have used these raw materials to produce the text above. The text has not been edited in any way.

　　I read the text written in my style by GPT-3, I have mixed feelings. On the one hand, I don't think these words are very good. GPT-3 copied some of the words I wrote and said myself, and combined them with various claims collected from the Internet. Among these claims Many I disagree with, and even some I think are unconvincing or simply ridiculous. The resulting article is a hodgepodge of words and ideas. I'm not worried at the moment - GPT-3 won't take away my job, at least Not in the next few years.

　　On the other hand, I was genuinely surprised. Reading the text above, I was dumbfounded. Is this written by artificial intelligence? ! Yes, the words are a hodgepodge of bits and pieces here and there. But isn't that the case with all the articles in the world? When I wrote A Brief History of Humanity, I also read many books, articles, and interviews, merging different ideas and facts into something new. The most shocking thing about the text produced by GPT-3 is that it actually makes sense. It's not a bunch of random sentences, but a coherent logic. I disagree with some of GPT-3's arguments, but to my horror, the text really makes an argument. If I give the monkey in "The Infinite Monkey Theorem" a typewriter and let it play, I only need to glance at it and within half a second be able to decide that what the monkey is typing is not my writing. However, I had to read the text of GPT-3 carefully for a minute or two before I could come to the conclusion that it wasn't my work.

　　There was something else that surprised and alarmed me. That is the pace of change. GPT-3 is still very primitive as artificial intelligence. Great is still to come. When I wrote A Brief History of Humanity in 2010, I hardly thought about artificial intelligence. It doesn't sound like a serious history book, but more like the subject of a sci-fi movie. But while I'm sitting in the library, immersed in old papers, my colleagues in the computer science lab aren't idle either. In 2011, when "A Brief History of Mankind" was first published in Hebrew, an artificial intelligence named Watson beat a human contestant on Jeopardy, a popular American quiz show, with a car. Self-driving cars roamed the streets of Berlin with the voice assistant Siri installed on iPhones.

　　10 years later, the AI ​​revolution is sweeping the world. This revolution marked the end of human history as we know it. Over tens of thousands of years, humans have invented various tools to make humans more powerful. Axes, wheels, and atomic bombs gave humanity new powers. Artificial intelligence is different. For the first time in history, power may be out of human hands.

Yuval Harari

　　All previous tools have made humans stronger because the tools themselves cannot determine their purpose. The power to make decisions is always in the hands of people. An axe can't decide which tree to chop, an atomic bomb can't decide whether to start a war. But AI can. Today, when you go to the bank to apply for a loan, it is already the AI ​​system that decides whether or not to approve your application. When you apply for a job and submit your resume, it is likely that an artificial intelligence will review your resume and decide your fate.

　　All previous tools made man stronger because man understands tools and tools don't understand man. A farmer knows what an axe can do, but the axe doesn't understand the farmer's needs and feelings. Soon, artificial intelligence will know more about ourselves than we do. Will it continue to be a tool in our hands, or will we be its tool?

　　Humans are still more powerful than artificial intelligence for at least a few years. Specifically, we still have the ability to dictate the development and use of artificial intelligence and other revolutionary technologies. We must use this ability wisely. The purpose of technology is never predetermined. The same technology can have many different uses. In the 20th century, some societies created totalitarian dictatorships through the power of electricity, trains, and radio, while others created liberal democracies using the exact same technology. Whether the new technologies of the 21st century are used to build heaven or hell is up to us to choose.

　　To make informed choices, we need to understand the full potential of new technologies, and we need to understand ourselves better. If we could use artificial intelligence to create a new world, what would that world look like? If you could use bioengineering to reshape the human body and mind, what would you change?

AI creation system Disco Diffusion: As long as you enter descriptive keywords, and then hand it over to the system for rendering, you can get a painting with great visual impact. The picture shows Van Gogh's "Starry Sky" and "Sunflower" rendered by the system.

　　In fairy tales, when a magical goldfish or an omnipotent elf promises to grant three wishes, it usually doesn't work out well. What people ask for is wrong because they do not know the true source of their suffering or happiness. If you want a better ending than the clumsy people in fairy tales, you need to know what it means to be human. who are we? Where are we from?

　　Fortunately, as the power of digital devices has increased exponentially over the past decade, so has our knowledge of human biology and human history. Since the publication of A Brief History of Humanity in 2011, scientists have discovered several new branches on the human family tree. Bones belonging to Homo naledi were first discovered in South Africa in 2013. In 2019, Homo luzonensis, who once inhabited the island of Luzon in the Philippines, was discovered. In 2021, two possible new human species were discovered in Israel and China, respectively.

　　Not only do we learn about the existence of other ancient peoples, we also learn more about their lives—the food they ate, the way they behaved, and even who they copulated with. When I wrote A Brief History of Humanity, we had only a few small clues that suggested interbreeding between Homo sapiens and Neanderthals. Now, we have a lot of evidence that they met and intersected. We also know that both Homo sapiens and Neanderthals interbred with another mysterious human species called Denisovans, first discovered in 2010. Recently, scientists discovered the skeleton of a mixed race whose mother was Neanderthal and father was Denisovan.

Bones of Homo naledi

　　Since I wrote A Brief History of Mankind, numerous details and new twists have been added to the human story. However, my central message in the book remains unchanged: Homo sapiens is best described as a storyteller. We create fictional stories about gods, nations, and corporations that form the foundation of our society and the source of meaning in our lives. We know more truth than any other animal, but we believe more fiction. But only humans kill each other for stories, and my thesis is that if you really want to understand human history, you have to take our fictional stories seriously. Just looking at economic or demographic factors is not enough.

　　Just take World War I as an example. Why did Germany and Britain go to war? Not because the territory isn't big enough or food isn't enough. In 1914, the two countries had enough territory to build houses for all their citizens and enough food to feed them. But the two countries couldn't agree on a common story that they could both believe, and they went to war. Today, Britain and Germany are at peace, not because their territories have increased (actually they have much less territory than they did in 1914), but because they now have a common ground that most Britons and Germans believe in s story.

　　The Buddha said thousands of years ago that the human world is a dream. In the age of artificial intelligence, the stories we believe in are more important than ever, because we have more powerful technology for pursuing fantasies.

　　When the ancients imagined heaven and hell, their fantasies deeply influenced their behavior. However, they can only postpone heaven to an imagined afterlife. In the 21st century, at least some people will be tempted to use artificial intelligence, bioengineering and other revolutionary technologies to try to live out their fantasies in this world. If we don’t choose what we believe in carefully, we risk being misled by naive utopian fantasies and plunged into technological hell with nowhere to go.

　　To create a better world, it is not enough to learn how to code or decipher the genetic code for computers. Artificial intelligence and genetic engineering could easily be used to serve totalitarian tyrants and religious fanatics. What we really need to understand is the human mind and the fantasies that the mind generates and believes. This is the task of poets, philosophers and historians, and it is more urgent than ever.

　　I am often asked why A Brief History of Humankind resonates with so many readers. I think this book is successful because it fills a real need to tell a story about the latest scientific discovery in human history. However, I also try to help readers see the big picture from a new perspective. We live in a global world, but what is taught in most schools and in books is the local history of a country or culture. To understand your future, you need to understand the history of the entire world and the challenges facing all of humanity. I wrote this book to help people see the world more clearly and empower everyone to participate in the most important debates of our time.

　　I don't fully agree with the words written by GPT-3 pretending to be me. But there's one sentence in it that I totally agree with: "We live in a prison of ideas invented long before we were born by people very different from us." It seems to me that the job of historians like me is to show these How the prison of thought is constructed and shows that we are not doomed to be locked in and unable to escape. Humans are storytellers; human society cannot function without stories. But at the end of the day, these stories are just tools we create to help people. If a story does more harm than good, we can always change it.