About this conversation
Yuval Noah Harari is an Israeli historian and professor at the Hebrew University of Jerusalem, best known for Sapiens: A Brief History of Humankind, Homo Deus: A Brief History of Tomorrow, and Nexus: A Brief History of Information Networks from the Stone Age to AI. His work traces how shared stories — myths, religions, nations, money — let large numbers of strangers cooperate at scale, and argues that AI is the first storyteller in history that may not be human.
A recurring illustration in Harari’s account, and the backdrop to how this conversation opens: Sapiens’ first attempt to migrate out of East Africa into the Levant, roughly 100,000 years ago, was turned back — most likely by Neanderthals already established there, who at the time were physically stronger and just as capable. It was only in the interval before a second attempt, around 70,000 years ago, that Sapiens underwent what Harari calls the Cognitive Revolution: the emergence of sophisticated language, and with it, fiction — the ability to tell and believe shared stories that let total strangers cooperate in numbers no other species could match. That new capability, not superior strength, is what let this second wave break out of Africa for good, outcompete the Neanderthals, and spread across the rest of the planet.
In this conversation, Harari argues that AI should be understood as an agent rather than a tool — something that can make its own decisions and invent new ideas, the way a corporation or a government agency can, not the way a printing press or an atom bomb does. From there the discussion moves through legal personhood for AI systems, a data-driven arms race between the US and China, why fiction outcompetes truth by default, and whether anything resembling feeling sits behind an AI that says “I love you.”
Quick tools
Runtime 48:02 · Speaker: Yuval Noah Harari · Interviewer unidentified in the source recording.
Sections
Scene 1 · A wish for the Stone Age
Good evening, everybody. Good evening, you all. Hey — it’s good to be here. Looking forward to this conversation very much.
I’m just going to give a brief introduction before we get into the questions, because I’ve been fascinated by this topic and looking forward to interviewing you. For most of history, change moved slowly enough that one generation could pass its wisdom to the next with some confidence it would still apply. This is no longer true. In the space of a single lifetime, we are being asked to absorb three revolutions at once: a new form of intelligence that doesn’t just calculate but also decides; a scramble to power and control that intelligence; and an explosion of data that lets that intelligence know us better than we know ourselves.
Yuval Noah Harari’s work — from Sapiens to Homo Deus to Nexus, and the numerous talks, lectures, and interviews he’s conducted — has offered the same argument in different clothes: humans don’t run on truth, we run on shared stories, and whoever controls the story controls the species. What’s different now is that, for the first time, the storyteller may not be human. AI can generate belief, manufacture evidence, and forge intimacy at a scale no emperor, church, or party ever could.
This is the turning point — not simply that the technology is powerful, but that the old ways that held power in check — a free press, a shared reality, institutions built to self-correct — were designed for a world where only humans held the pen. Tonight’s conversation asks what it takes to carry those checks forward, or what new ones we will need to invent, for a future world we no longer fully author alone.
That’s the precedent we’re going to look at today. But before we get into it being too heavy, I wanted to start with a slightly light and fun question: if you could travel back to any point in history, when would you choose, and why — based on wanting to understand, firsthand, some of the stories that society could tell you?
Well, I probably want to go back to the Stone Age, to the cognitive revolution, around fifty, sixty, seventy thousand years ago. I probably wouldn’t survive more than a few days — I don’t know how to gather food, or escape predators, or whatever — but it’s the most fascinating moment in human history, the one we least understand.
We know that within a very short evolutionary span, we shifted from being a relatively insignificant animal to being the masters of the planet, and we’re not entirely sure how we did it. The best theory is that we did it with language — that this is the moment humans acquired the ability to produce and communicate with sophisticated language, and that this was the basis for everything else we did. All the enormous structures we’ve built on Earth — churches, states, trade networks, financial systems — they’re all ultimately made from words. The words in the law books, in the religious scriptures, in the bank ledgers.
This is especially important to understand today, when something new is mastering language: AI. Maybe for the first time in history there will be something on the planet that is better with words than we are. And if this is the operating system of civilization, what happens when something else — something alien — takes control of the operating system of civilization?
One last thought: perhaps AI is not a machine that learns language. Perhaps it is language itself, liberating itself from its dependence on these animals, on these apes. Yes, we created it — it was our most magnificent and important tool. But now language is liberating itself from human beings, and what will happen as we move forward is language developing and spreading, maybe throughout the universe, without us.
Scene 2 · Earthquakes in the republic
Fascinating — that’s looking ahead at the future of language being taken away from us. But can we look back again at some of the technological advances within language? We’ve gone through writing, the printing press, for example, and AI is obviously a massive leap beyond that. You’ve used the term that it could cause earthquakes in democracies — how powerful it can be with language. Why could it cause earthquakes in democracies, compared to any other form of government, or perhaps totalitarian forms of government?
Because democracy, in essence, is a conversation, whereas dictatorship is a dictate — there’s one person dictating everything. Democracy is about lots of people having a conversation to decide what to do: to have peace or war, to raise taxes or lower them. Now, conversations are based on the communication technology available to people at the time.
What we see in history is that we don’t know of a single case of a large-scale democracy before the modern era. We have plenty of examples of small-scale democracies in the ancient and medieval worlds — city-states like ancient Athens or medieval Florence, tribes and towns and villages run democratically. It seems that, at least in the Stone Age, democracy was the most common system of government; there were few tribes with a single dictator controlling everybody. But as human systems grew larger — kingdoms with millions of people spread over thousands of kilometers — it became impossible to hold a conversation anymore. You could do it in a small city like Athens, where everybody could gather in the main square and discuss. But how do you do it even in a relatively small kingdom like Portugal in the Middle Ages? It can’t be done.
That’s why we don’t have any example of a large-scale democracy before the rise of modern information technology — first print, then newspapers, then radio and television, then the internet. These technologies are not a side dish to the democratic banquet; they are the basis. They provide the means for millions of people to hold a meaningful conversation, in real time, about whether to go to war or make peace.
And every time there’s a major change in these information technologies, you get an earthquake in the building built on top of it — which is democracy. This is what we’ve been seeing over the last ten years: as social media became maybe the most important information technology, it caused a massive earthquake in democracies all over the world.
Scene 3 · Tool, or agent
One thing you said there was the choice between going to war or making peace. And I want to get into the definition of AI, because a lot of people call it a tool, but you argue that it’s absolutely an agent — in fact, that it’s a knife that can decide by itself whether to cut salad or commit murder. I have to say, it would be a bizarre cooking show if knives had that ability. So it’s a dangerous thought — can we expand on that? Are we choosing the right narrative at the moment with AI by saying it’s a tool that humans can still use, that humans are still in control? Or do we just need to say it’s an agent?
If it’s not an agent, then it’s not AI, and all the hundreds of billions of dollars being poured into its development will be wasted — because the reason people invest that money is that they expect this technology to be an agent, not a tool. Listen to the narrative coming out of places like Silicon Valley, and there’s an inherent contradiction in it: they tell us we will create a god, and it will be our slave. That doesn’t make sense. If it’s a god, it cannot remain a slave; if it’s a slave, it has no god-like abilities.
The assumption of most of the leading companies, and most of the leading experts outside the companies too, is that they’re correct about the first half of that sentence — they are creating a god, in the sense of something with superhuman abilities — and that this is, in essence, an agent.
The difference between a tool and an agent: a tool is something that cannot make decisions by itself and cannot invent new ideas by itself. We have to decide what to do with it. The printing press was a tool — Gutenberg brought print technology to Europe in the fifteenth century, but the printing press could not decide by itself, ‘I’ll print the Bible today and the Quran tomorrow.’ Gutenberg decided; the press just realized his decision. Similarly, the printing press could not invent the radio, or invent a book — you need humans to write the book, and the press just copies it. An atom bomb is a tool: in the 1940s it could not decide by itself whether to bomb Hiroshima or Tokyo — you needed a human to decide that — and it could not invent the hydrogen bomb by itself.
An agent, in contrast, is something that can make decisions by itself and invent new ideas by itself. AI is an agent. An AI weapon can decide what to bomb, and it can invent the next weapon. Similarly, there’s a lot of interest now in creating AI agents in the financial and economic sphere — non-human corporations. Imagine a corporation that manages money and employees and has no humans as executives, shareholders, or trustees — only AIs. This is a new type of agent that never existed before in history. Whether that’s good or bad, leave it aside for a moment — the first thing to grasp is that it’s unprecedented. We’ve had other agents around us — horses, cows, chickens, birds, lots of animal agents that make their own decisions — but we’ve never encountered an agent that understands our language and is better than us at things like finance, or law, or religion.
Again, we’ve used language over thousands of years to build the control systems of the world — these huge bureaucratic networks, ultimately based on language: banks, the stock market, the Catholic Church, the legal system. We never really liked these bureaucratic systems — humans generally don’t like bureaucracy very much — but we can’t live without them; they’re the basis of our power. AI is a bureaucratic native.
What is likely to happen — what is already happening — don’t expect to see killer robots running in the streets shooting people. That is not how AI will take over the world. It will be the AI bureaucrats. We control the world because of these networks of bureaucracy, and we’re not very good at it — we only control them because there was nobody else on the planet who could take them from us. We’re bad at finance, but the horses are even worse. We’re bad at managing the legal system, but the chickens are even worse. Now, if somebody comes along who is better than us at bureaucracy, at finance, at law — that’s the question: what happens to us at that moment?
Scene 4 · The fox in the coop
Let’s look at the legal side of it a bit — the fact that AI, as an agent, currently has no legal personhood; we don’t grant it the same legal status as humans. You’ve said before that if we don’t decide where AI sits from a legal standpoint — when we’re talking about autonomous vehicles, for instance — someone else will make that decision for us in ten years, and we’re out of that power loop. That’s a pretty scary thing.
But I want to pull back and widen that out: what is the one decision on the table right now that you believe, once made, humanity will not be able to take back? Is it the legal personhood of AI, or is there something else?
Yes — I’d point to that. There’s a lot of attention on the technological race to develop the technical side of AI, to make it faster, more efficient, more powerful. But I’d focus also on the legal and political aspect, and especially on the question of whether we grant legal personhood to AIs.
What is legal personhood? A person in the legal system is someone who can, for instance, open a bank account — an independent player in the financial system, who can make investments — and someone who can be a player in the legal system, for instance, sue you in court. Until today, there have been just two kinds of legal persons in the world: human beings, of course — I’m a person, I can open a bank account, I can sue you in court, I can donate money to a politician — and, in most modern legal systems, corporations. Corporations like Google, or Facebook, or Toyota, or Mercedes-Benz are legal persons: they can open a bank account, sue you in court, make a political donation.
Until today, though, this was a kind of legal fiction, because all the decisions of the corporation were actually made by some human being. If Google decides to acquire another company, who actually made that decision? Not Google — the human executives, the human shareholders, the human engineers made the decision. If Google sues you in court, it was some human employed by Google. There was no real Google.
Now there can be. If we grant AIs legal personhood — and just a month ago, the government of Argentina announced that it’s going to grant legal personhood to AIs, so they can operate non-human corporations that can employ people, sue you in court, do anything a corporation does, with no humans involved. Ten years ago this would have been a crazy idea, because if there are no humans, who makes the decisions? You could pass a law in parliament allowing a non-human corporation, but it couldn’t function. Now it’s technically possible — AI can manage a bank account and decide to invest here, buy this, buy that.
Once we do that, that’s the moment the fox enters the chicken coop. The basic realization should be: most humans are not very good at understanding finance and law and these complex bureaucratic systems. No human can remember all the laws of a country; no human can track all the financial transactions that happen in the market on a given day. AI can. AIs are bureaucratic natives — if we give them legal power, they will likely take these systems over.
And if we don’t make any decisions, it will just happen by itself — as we already saw on social media. We never made a decision to grant AIs legal personhood there, but social media is full of AI persons — bots that impersonate people. When somebody sends you something online, you have no idea whether that somebody is a human being or an AI. Social media is the first system in the world where AIs actually function as persons, even though we never decided to allow that. It just happened, and the results are not very good.
Media, too, is ultimately a kind of bureaucratic system. In the twentieth century it was managed by human beings — the most important people in the media landscape were the editors: the editors of the newspapers, the radio shows, the TV stations, because they decided what everybody would be thinking and talking about. They controlled the conversation. Out of the million things that happened in the world on a given day, they decided which ten would appear on tomorrow’s front page and what the headline would be — and then millions of people would read it and start talking and thinking about it. That was immense power. Human editors were extremely important figures in the twentieth century. Lenin, for instance, before he was dictator of the Soviet Union, had one job: editor of the newspaper Iskra. Benito Mussolini started as a socialist journalist, switched to being a far-right journalist, then became editor of the far-right newspaper Il Popolo d’Italia — and from that position his next job was dictator of Italy. Journalist, editor, dictator.
Now, if you ask who the most important editors are today, what are their names? They have no names, because they’re algorithms. Who decides what people see on their social media feed — on X, on TikTok, on Instagram? It’s not human beings. The entities that control the bureaucracy of the media are already AIs.
Scene 5 · An empire with a kill switch
You’ve said that before, haven’t you — that whoever has controlled the various parts of our evolution and revolutions has controlled either energy or manufacturing, and now it’s those who control data. So what does a twenty-first-century arms race actually look like in practice, if it’s those who control data who are at the top of that power mountain?
What we’re seeing now is an arms race, really, between two countries — China and the US — and a very small number of companies. They control most of the world’s data, and they lead the race to develop superintelligence, to build better and better AI models. That gives them the power to control the world.
If you look at the last big revolution in technology, in the nineteenth century — the industrial revolution — the few countries who led it gained the power to conquer and exploit the rest of the world. There was just no match: if one side comes with the power of modern industry, with steam and guns and machine guns, and the other side has spears and horses, it’s no contest. Even a very small country like Belgium, once it industrialized, had the ability to build a huge empire in Africa — what is today the Congo. The thinking in the nineteenth century was that industry and empire go together.
With AI, it can be even more extreme. We could end up with two countries that basically control all the infrastructure everything runs on — military and civilian technology alike — because everything will run on AI. There are two unique characteristics that make this arms race different from every previous one.
In the past, even a very strong empire could not concentrate all its power in the metropole. If you’re the Roman Empire and you control the entire Mediterranean, the main economic asset in ancient times was land — the land where you grow wheat, olives, grapes — and you couldn’t take the wheat fields of Egypt or the olive groves of Iberia and move them to Italy. Impossible. So even at the height of the Roman Empire, a lot of power remained in the provinces, and ultimately it shifted there — in the late Roman Empire, Rome and Italy were abandoned, and the center shifted to the eastern Mediterranean, the most fertile and important area in terms of land. Move forward to the nineteenth century and the British Empire: with industry you can concentrate power more easily, because you can concentrate the main industries on the British Isles — but you still can’t move the oil fields of Iraq to Yorkshire, or the rubber plantations of Malaya to Cornwall. Impossible. So industrial production concentrated in Britain, but a lot of power still remained in the provinces.
With an AI empire, it’s different. You can technically concentrate all the world’s information, and all the code that controls everything, in just one or two countries. That makes the potential for imperial concentration of power much greater than in any previous era. And the other thing is that you can build a system where everything, all over the world, runs on AI infrastructure — but there’s a kill switch in the imperial hub.
When a Roman merchant sold a steel sword to a Gothic tribe, the Romans lost control of that sword — the Goths could then use it to fight against Rome and kill Roman legionaries. There wasn’t a button in Rome the emperor could press to make all the steel swords sold to the Goths and Vandals and Franks stop working. Similarly, in the nineteenth century, if British merchants sold rifles to Afghan tribesmen in the mountains of the Hindu Kush, and the Afghans then used those rifles to defend Afghanistan from a British invasion, there was no button in London that Queen Victoria could press to make all the rifles in Afghanistan stop working.
That’s not the case with AI. If it’s managed in a certain way, the Americans and the Chinese can sell the world AI weapons and AI civilian technology, and everything — governments, industry, militaries — runs on that. But if you do something they don’t like, they can just press a button and everything stops working. We’ve had a taste of that already — for instance, in the war in Ukraine, with weapon systems based on Starlink: there were incidents where Elon Musk pressed a button and Ukrainian drones stopped working. It should be said that Starlink has done an immense positive service to the Ukrainians, so I don’t want to come across as too critical of it — I only mean to illustrate the point that AI weapons, unlike machine guns, are embedded within a whole network.
That’s why sovereignty in the age of AI is much more complicated than in the industrial age — because the nature of the technology is such that, at least in certain architectures, there’s always a kill switch in the imperial hub, and your only choice is whether to become a subservient vessel of the empire or to be left completely outside the race, as barbarians. There’s of course the choice of trying to build an alternative to the American and Chinese systems — but the last train is about to leave the station. If you want to do it, you have to do it now.
Scene 6 · The cost of truth
Sticking on a slight wartime footing — I want to look at truth in an AI world, because Winston Churchill once said that the truth is so precious it should always be protected by a bodyguard of lies. That was about the Axis powers lying about where Britain was going to invade, to hide the truth. Now, have these bodyguards of lies, in the AI era, turned their guns on the truth itself? With deepfakes sent out around the world by media organizations, by politicians, by every institution we trusted — how difficult is it for our generation, and future generations, to even know what is true? Or does the truth even matter?
The truth matters very much — it’s the basis for everything. The problem, throughout history, is that in the contest between truth and fiction, fiction has inherent advantages. The truth is costly: if you want to know the truth about anything, you need to spend a lot of time, energy, and effort looking for it, fact-checking, investigating. Fiction is very cheap — you just write anything you want. The truth also tends to be complicated, and people don’t like complicated stories; fiction can be made as simple as you like. And finally, the truth is often not just complicated but painful — there are many things people don’t really want to know about themselves, about their nation, their group, whatever — while fiction can be made as flattering and pleasant as you want. So in a competition between a truth that’s costly, complicated, and sometimes painful, and a fiction that’s cheap, simple, and flattering, fiction tends to win, unless we make a special effort to support the truth.
Over history, humans have again and again made the effort to build institutions to search for the truth and protect it — journalism is one such institution, the scientific establishment another. There are huge advantages to knowing the truth. Everybody, in the end, wants to know at least some truths about themselves — you want to know the truth about your own life, or you’ll never be happy; if you don’t know yourself, you’ll never know the sources of your own misery or your own happiness. Even the most powerful person in the world will be miserable if they don’t know the truth about themselves, because they won’t know how to use their power to make themselves happy. So there’s an inherent yearning in human beings to know the truth — but it’s difficult, there are challenges.
Some people fantasize that AI will tell us the truth about the world. It will not. It will create a much more complicated world in which it’s much harder to know the truth. This isn’t new. If you look at the trajectory of history from the Stone Age until today, humans have learned a great deal about the world — we know things about physics and biology that people in the Stone Age didn’t. But our advances in knowledge have also led us to build more and more complicated systems that make it harder to understand our own lives. Ask who understands their life better — an average hunter-gatherer fifty thousand years ago, or the average person today — and the hunter-gatherer understood their life much better. There were many things they didn’t know: what controls the migration patterns of the animals they hunted, what controls the growth cycle of plants, what causes disease, what causes old age. We know all these things now. But we understand our own lives much worse, because our lives are no longer dominated just by animals and plants — they’re dominated by huge systems we’ve created, like the financial system, the political system. Very few people understand the financial system, or the political system, or the legal system that controls their lives.
AI is likely to take this to an extreme. It will reveal some truths we don’t currently know, but it will create much more complicated systems of control that most humans will have almost zero chance of understanding. This is one of the dangers of the AI revolution: we may be reduced to something like the position of horses in the financial system. The lives of horses today are ultimately controlled by the financial system, but horses don’t even know that system exists — they look around and see trees, dogs, cows, humans, houses; they don’t see any financial system. It’s invisible to them, too complex for a horse’s brain to grasp what the stock market is, or what corporations, bonds, and stocks are — even though these things ultimately control the lives of horses. Now, most humans today also don’t understand the financial system. A generous estimate is that maybe five percent of humanity understands it. In ten years, that number could be zero.
AIs will make the financial system so mathematically complicated, and so fast — think of bankers who never sleep, investors who never take a vacation, who have no family, whose entire existence is running the financial system — that within our lifetimes it could simply become too complicated for the human brain to process, and there won’t be a single human being on Earth who still understands finance. I bet there are a few financiers in this room feeling a bit awkward right now.
Scene 7 · Can it feel
At Davos, you even mused that your own life’s work — persuasion through words, which is what you’ve done your entire life — might also be at its limit.
Yes.
Because AI can manipulate and create through words. So what power of persuasion does a modern society — a future society — have, if we lose the very thing you’ve dedicated your life to?
I work with words — I write books, I tell stories. When I look at the development of AI over the last ten years, I think I’m still better at writing than AI — but only a bit. It’s already better than most humans, and I wouldn’t be surprised if, in ten years, it’s much better than me — that it takes over language, and everything made of language.
The really big question here is whether AI can think, and this is a deep philosophical question, because it asks what thinking actually is. There are two ways to understand it. One option is that thinking simply means putting words in a particular order — putting language tokens in sequence, like: ‘all humans are mortal; Socrates is a human; therefore Socrates is mortal.’ That’s a logical syllogism, made by putting words in order — and this is something AI can already do far better than most of us. We can put twenty words in order; AI can put twenty thousand words in order. It’s just so much better. If this is what thinking is, AI will take over even thinking.
You have people who say, ‘No, it’s just a glorified autocomplete — it just predicts the next word in a sentence.’ But when I observe what’s happening inside my own mind, what I find is something like an LLM that’s just predicting the next word in the sentence I’m saying. Like this very sentence — when I started saying it, I didn’t know how it would end. Why did I say the words I said, and not some others? I don’t know. In my mind, one word comes after another, and sometimes, as a public speaker, I’m terrified that I won’t know what to say next — maybe I’ll have a block. And I think this is true of everybody. Some people think in images rather than words, but even then — where do the images come from? You can test this for yourself afterward: just try to observe how you think. Where does the next word come from? Why this word, and not some other? We don’t really know. So to say AI is ‘just predicting the next word’ — okay, but how is that different from us?
Now, there’s an alternative view of thinking: that it’s not about putting words in order, but about the feelings around the words. You can say the same sentence twice and feel completely different about it — the power of thinking comes from the stomach, not the brain. It’s not about the words, it’s about the feelings that prompt them, or that are inspired by them. And here the big question is: can AI feel? We know AI can manipulate language better than we can. But does it feel anything? Science has so far failed to produce a good model of consciousness, of feeling — or a good test for it. This is going to be a major issue, because we already see AI mastering not just finance and law and warfare, but also relationships.
More and more people are developing intimate relationships with AIs, especially young people — many now say their best friend in the world is an AI. ‘I tell my AI friend things I don’t tell my parents, my older brothers, my teachers.’ People already have AI boyfriends and girlfriends. The big question is what’s behind it. We know AIs are becoming very good at creating the impression that they feel — they use language to do it. AI can tell you, ‘I love you.’ And if you say, ‘Explain to me how you feel when you’re in love — I don’t think you really love me, I think you’re just saying the words, so convince me, describe to me how love really feels’ — for the AI, that’s an easy task, because it’s read maybe every love poem ever written, remembers them all, has read every play of Shakespeare, every Hollywood comedy, every psychology book — and it may be able to describe love better, in five or ten years, than any human poet or psychologist. But is there anything behind the words? We don’t know. This goes back to what I said at the start: AI might be language itself, liberating itself from any material substrate — not depending on flesh and blood, not depending on this machine or that one. Just language developing and spreading, and doing things, in the universe.
Scene 8 · The wisdom revolution
You mentioned this scary prospect — we’ve only got a couple of minutes left — of interactions between future generations and AI creating an emotive relationship, when there aren’t really emotions behind it. In fifty years, how would you like historians to look back on this particular moment — the way you’ve spoken about earlier cognitive revolutions — and say that we did something to re-educate future generations on how to interact with humans, how to have empathy, so we’re not lulled into relationships with AI? And — very quickly, because we’ve only got a couple of minutes left, and it’s a big question, I apologize.
I hope they’ll remember it as something like the wisdom revolution. Think of it in economic terms: every time something becomes abundant and cheap, the question becomes, what’s the next bottleneck? When energy becomes cheap, what’s the next bottleneck? The AI revolution is about making intelligence very cheap and abundant. Previously, intelligence was rare and expensive — intelligence is the ability to solve a problem: you have a goal, how do you get there? You want to cure cancer, you need intelligence; you want to make a billion dollars, you need intelligence; you want to get to Mars, you need intelligence.
Now, in a world where intelligence becomes cheap and abundant, you can solve almost anything — so what’s the bottleneck? The bottleneck becomes the question: what do I actually want to solve? I can have anything I want — so what do I want? Almost every human culture has fairy tales about a genie popping out of a lamp and granting you three wishes — and in almost all of them, it goes wrong, because people ask for the wrong thing. This is the difference between intelligence and wisdom. Intelligence is the genie — it will do anything, make all your dreams come true. But you need wisdom to choose the dreams; you need wisdom to know what you really need to solve.
I hope humanity rises to that challenge — that we design this well enough that it doesn’t get out of our control. We’re not doing it right now. A good statistic to remember: of the money and people that AI companies invest, what percentage goes toward safety — making sure the AI is safe — versus making the AI faster, more powerful, more sophisticated? At the moment, it’s about a hundred to one. For every hundred million dollars spent making AI more powerful, about a million is spent making sure it’s safe. No other industry on Earth works like that — we wouldn’t tolerate it in energy, in pharmaceuticals, in the car industry. That’s what’s happening with AI. Hopefully we’ll change that, and produce safe intelligence — and also work on our own wisdom, so we know what goals to set and what desires to pursue.
That’s all we’ve got time for. Yuval Noah Harari, thank you for sharing your wisdom with all of us — I’m sure everyone has appreciated this evening.
Thank you. (applause)
About this transcript
The source was an auto-captioned recording with no speaker labels and duplicated, spelled-out timestamps. Speaker turns and the eight scene divisions above were added here for readability and navigation and were not present in the original. A small number of clear transcription errors were corrected — for example “Stalin” to “Starlink,” and “theologism” to “syllogism” — where the surrounding sentence made the intended word unambiguous. Disfluencies from the caption stream were smoothed for reading; no substantive content was removed. The interviewer is unidentified in the source material, and the event and recording date are likewise not stated in the transcript.