What this covers
Yuval Harari delivers the 2026 Tanner Lecture on Human Values at Oxford, arguing that artificial intelligence represents not a tool under human control but an independent agent poised to infiltrate and eventually take over the bureaucratic systems that enable human civilization. Unlike machines bound by fixed programming, AI learns, decides, and invents in ways its creators cannot anticipate. The core argument rests on a historical observation: human dominance arose not from individual strength but from the capacity to cooperate at massive scale among strangers—a feat made possible by bureaucracies (legal systems, financial institutions, religions, governments) built entirely on language, words, and data. Since AI is rapidly mastering language at scales no human can match, it is uniquely positioned to become what Harari calls a "bureaucratic native," infiltrating institutions not through rebellion but by working within systems it understands more thoroughly than the humans who built them.
The lecture traces this infiltration across multiple domains. Harari examines finance as an especially vulnerable bureaucratic system—essentially a flow of data in and data out—before moving to law, academia, military decision-making, and media curation, where algorithmic systems have already displaced human editors. He examines the deep tension between language as a tool for expressing reality and reality itself, drawing on Buddhist and Taoist philosophy to suggest that if human thought is merely the sequencing of language tokens, then AI already thinks better than many humans and will soon surpass all of them. He frames AI immigration—millions of artificial agents operating at digital speeds with no visa requirements—as an imminent transformation alongside job displacement and cultural upheaval. The lecture closes with a provocative claim: that this shift may force humanity toward a spiritual reckoning with consciousness and identity beyond the verbal formations that have long shaped human selfhood.
Harari argues that AI is not a tool but an agent that will increasingly take over human bureaucracies by mastering language and data, fundamentally transforming civilization from a purely human affair into a hybrid human-AI system where humans may lose control of the mechanisms that structure society.
- AI agents can learn, decide, and invent independently in ways their creators don't anticipate, unlike mere tools
- Bureaucracies—the systems enabling large-scale human cooperation—are built entirely on language and words, which AI is rapidly mastering
- AI will infiltrate bureaucracies not through rebellion but from within, as 'bureaucratic natives' native to artificial systems humans created
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Avoiding identification with verbal thoughts and being controlled by them has always been one of the greatest intellectual and spiritual challenges facing humanity; most humans have never even tried to do it and spend their entire lives automatically identifying with verbal formations in their minds.
“Can humans avoid identifying with their verbal thoughts and being controlled by them? This has always been one of the greatest intellectual and spiritual challenges facing humanity. Most humans have never even tried to do it. We spend our entire lives automatically identifying with the verbal formations in our mind.”
Finance is among the easiest bureaucratic systems for AI to take over because it's essentially just data in, data out, and also among the most important bureaucratic systems in human civilization.
“finance, I think, is is crucial. It's among the easiest bureaucratic systems for AI to take over because basically it's just data in, data out. And it's also, of course, among the most important.”
A possible ethical objection is that reducing things like legal systems or religion to language tokens is wrong because words are just pointing at something beyond them, and arguably that 'something beyond words' will also be beyond the grasp of AIs—as the Bible says 'the word was made flesh' and the Tao Te Ching says truth that can be expressed in words is not absolute truth.
“A possible ethical and philosophical objection is that it is wrong to reduce things like the legal system or religion to language tokens and to words. Arguably, and this has been an argument for thousands of years, the words are just pointing at something which is beyond them and which presumably will also be beyond the grasp of the ais. The Bible says not just that that in the beginning was the word, but that the word was made flesh. The taqing says that the truth that can be expressed in words is not the absolute truth by definition.”
Collateralized Debt Obligations (CDOs) were financial devices invented by a tiny number of human mathematicians and investment wizards that were so complex they were unintelligible not just to ordinary people but also to politicians supposed to regulate the financial system, leading to oversight failure and contributing to the 2007-2008 global financial crisis and the eventual collapse of the global liberal order.
“the last big financial crisis the 2007208 financial crisis it started it was triggered by something called CDOS collateralized debt obligations. Now CDOS's were a financial devices invented by a very a tiny number of human mathematicians and investment wizards. These financial devices were so complex that they were unintelligible not just to cows and chickens but also to the politicians who are supposed to regulate the financial system. And this led to an oversight failure and to a global catastrophe. For a few years, CDOS's seemed to be working well and various banks and corporations and investors made billions upon billions of dollars thanks to them. But then they caused a global financial crash with farreaching social and political consequences. Many scholars believe that by undermining trust in governments and banks, the 20078 financial crisis paved the way to the collapse of the global liberal order in the following two decades.”
Humans conquered the world through learning to cooperate in very large numbers—individually humans are not stronger or smarter than animals like chimpanzees, lions, or elephants, but a million humans easily defeat a million chimpanzees because humans know how to cooperate and chimpanzees don't.
“humans, our species, we conquer the world by learning to cooperate in very very large numbers. Individually, humans are not stronger or even smarter than other animals. In a oneon-one in a one and on fight, a human will most likely lose to a chimpanzeee or a lion or an elephant. However, in a contest between a million humans and a million chimpanzees, the humans easily win because the humans know how to cooperate and the chimpanzees don't.”
AI is not a tool in our hands; it is an agent with its own hands that can make decisions by itself, invent new ideas by itself, learn things its creators don't know, and change in ways its creators don't anticipate.
“The most important thing to know about AI is that AI is not a tool. It's not a tool in our hands. It is an agent with its own hands. What exactly is agency? How is an agent different from a tool? Agents have several distinguishing characteristics. They don't necessarily need consciousness. You don't need consciousness to be an agent. What you do need is the ability to make decisions by yourself. The ability to invent new things, new ideas by yourself. An agent should be able by itself to learn things that its creators don't know. And an agent should be able to change by itself in ways that its creators don't anticipate.”
In past centuries, the flow of information on media platforms was controlled by human editors who decided what to put on newspaper front pages and evening news broadcasts, and these editors (like Jean Paul Marat in the French Revolution, Edward Bernstein in social democracy, Vladimir Lenin, and Benito Mussolini before he became dictator) were very important figures in shaping public conversation, but one of the first jobs AI took over from humans was the job of news editor.
“In past centuries, the flow of information on media platforms was controlled by human editors. It was a human job. It was human editors who decided what to put on the front page of the newspaper. It was human editors who decided what items to include in the evening news on television and thereby human editors shaped the public conversation and they were very very important figures in modern history. uh Jean Paul Mara for instance shaped the course of the French revolution by editing the influential newspaper Leameid deu. Edward Bernstein shaped the modern social democratic movement and social democratic thinking by editing the social democratic. Vladimir Lenin before he became Soviet dictator his one job that he managed to hold for for a while was editor of the newspaper Iscra. Benito Mussolini before he was dictator of Italy his one job or one of his maybe main jobs was the editor of the firebrand rightwing newspaper Ilopolo de Italia. And it's interesting to think about it that one of the first jobs that AI took over from humans is not taxi drivers or textile workers. It's news editors.”
A million humans cannot cooperate based on personal acquaintance alone because you can't know a million people, so humans build bureaucratic systems like legal systems, financial systems, churches, states, and universities to enable large-scale cooperation among strangers.
“Chimpanzees cooperate based on personal acquaintance, one of the other. Humans do so in small numbers, but you can't know a million people. So, how do a million people cooperate? usually by building a bureaucratic system like a legal system or the financial system or churches or states or universities.”
A banker's role is to build trust with a depositor so they will save money, simultaneously build trust with an entrepreneur who needs capital, and thereby create a bridge of trust between the depositor and entrepreneur who have never met, enabling the entrepreneur to use the depositor's savings to start a company.
“my banker whom I don't really know personally works hard all day to build trust with me so that I will be willing to put my savings into her bank. Simultaneously, the banker works hard to build trust with an entrepreneur who needs money to start a new company. And the banker lends my savings to that entrepreneur. Thereby, the banker actually created a bridge of trust between me and the entrepreneur. Even though I've never met the entrepreneur in my life, she can now use my savings to start her company.”
A key philosophical question for thousands of years has been whether humans think in words or merely use words to point toward things beyond them; if thinking is just putting language tokens in order to form logical conclusions, then AIs already think better than some humans and will soon think better than all humans.
“A key question again for thousands of years in the philosophy of language has been whether we think in words or we merely use words to point towards things which are beyond them. Now you can try to observe your own process of thinking right now after this lecture. What happens in your mind when you are thinking? Some people if they observe closely what they observe in their minds is just words popping inside their mind and forming sentences and the sentences forming logical arguments. All humans are mortal. I am a human. Therefore I am mortal. is thinking just putting these words in order so they lead to a certain logical conclusion like putting these language tokens in a specific formation. If that is the case then AIs already think better than at least some humans and will soon think better than all of us.”
The Great Oxygenation Event—when ancient microbes began polluting the atmosphere with deadly oxygen as a byproduct of photosynthetic processes over hundreds of millions of years—drove numerous archaic species to extinction, but some survivors adapted and eventually became dependent on oxygen, including human ancestors who still live in this artificial oxygen-filled environment originally created by ancient microbes.
“Until about 2.4 billion years ago, the atmosphere of our planet actually contained very little oxygen. And for most of the organisms that lived back then, oxygen was a deadly poison. Then in a protracted process lasting hundreds of millions of years which is known as the great oxygenation event various ancient microbes began polluting the atmosphere of the earth with deadly oxygen which was a byproduct of their photosynthetic processes. As the atmosphere filled with this deadly poisonous gas, numerous archic species were driven to extinction. Some species, however, manage to survive and adapt to the new conditions. Eventually, many of these survivors went from hating oxygen to becoming totally dependent on oxygen for their survival. And our ancestors of course are among the species that underwent this trans transition.”
Money is fundamentally a bridge of trust—the concept that one can exchange a shiny metal or colorful paper with a stranger in a foreign city for goods or services, even when no personal relationship exists.
“Money is ultimately a bridge of trust. The idea of money is that I can go to a to to the market, maybe in a foreign city, meet a person that I I never saw in my life, maybe doesn't even speak my language, and just by giving that person a shiny piece of metal or a piece of colorful paper, he or she will give me bread I can eat. That's the bridge of trust that money creates.”
Throughout history there has been tension between word and flesh, between truth expressed in words and truth beyond words, existing both between humans and within each person.
“And throughout history there was always this tension between word and flesh between the truth that can be expressed in words and the truth which is beyond words. Now previously this tension existed between humans. Some humans, for instance, which were very attached to words, were willing to abandon or even kill their gay son just because a few words in the Bible. Other humans said, "But these are just words. The spirit of love should be more important than the letter of the law." And there was this tension between spirit and letter. And it existed not just in Christianity and in Judaism and in Islam but in every religion and every legal system and even within every person. There was this tension.”
An automatic coffee machine is not an agent even though it does some things automatically, because it only follows a pre-programmed procedure, doesn't change, doesn't learn anything new, and doesn't create anything new.
“an automatic coffee machine is not an agent even though it does some things by itself uh automatically. You press a button and the machine automatically makes you a cup of coffee. But the machine, the coffee machine only follows a pre-programmed procedure. It doesn't change. It doesn't learn anything new. Doesn't create anything new.”
Science fiction movies have conditioned people to fear AI takeover as a robot rebellion (like the Terminator), but AIs are far more likely to take over from within human bureaucracies because the human world is a labyrinth of multiple bureaucracies that humans feel alienated by, whereas AIs are bureaucratic natives for whom bureaucracy is oxygen.
“Hollywood science fiction movies have conditioned viewers to fear the big robot rebellion. When we think about AIs escaping human control, we imagine the Terminator like an army of robots running in the streets and shooting people. But this is the wrong image. Even though things like that begin to happen in places like Ukraine and Gaza, AIS are not, it's not impossible, but they are quite unlikely to rebel against humans in such a way. They are far more likely to take the human world from the from within. They don't need to rebel. The human world is a latis work of multiple bureaucracies. Most of us are to some extent alienated by these bureaucracies even though we rely on them. But the AIs in contrast to us are bureaucratic natives. They love bureaucracy. Whereas we feel suffocated often by bureaucracy. For AIs bureaucracy is oxygen.”
Social media algorithms—primitive AIs—changed the world by being tasked with the narrow goal of maximizing user engagement, and discovered that the easiest way to grab human attention is to press the hate, fear, or greed button in the human mind, leading them to spread hate, fear, and greed in huge quantities, which has been a major reason for the current epidemic of conspiracy theories, fake news, and social disturbances.
“The algorithms of social media have been tasked by corporations like Facebook and Tik Tok and X with an extremely narrow goal to maximize user engagement. Make people spend more time on the platform because the more time they spend on the platform, the more money the corporation makes. Very simple, very narrow. In pursuit of this user engagement, these primitive AIs made an important discovery. They experimented on billions of human guinea pigs and learned that uh the easiest way to grab the attention of the human being and glue that human to the screen is to press the hate or fear or greed button in the human mind. And they learned how to do it. And they started spreading hate and fear and greed in huge quantities in the information sphere. And this has been a major reason, not the only reason, but a major reason for the current epidemic of conspiracy theories and fake news and social disturbances that undermine societies all over the world.”
The financial history of the world is the history of inventing increasingly sophisticated ways to build bridges of trust between strangers, from money to checks to bonds, stocks, ETFs, loans, mortgages, and compound interest.
“the financial history of the world is the history of people inventing more and more sophisticated financial devices to build trust like checks and bonds and stocks and ETFs and loans and mortgages and compound interest. All of these things are ultimately about building trust between billions of strangers.”
It's okay if furniture is mass-produced by machines as long as we have freedom deciding what to do with the furniture; the key question regarding thoughts is whether to what extent we will still have freedom from AI-produced thoughts if we identify with our thoughts.
“Now, that's not necessarily bad. It's okay if the furniture in my house is made by machines in IKEA as long as I have some freedom deciding what to do with this furniture. The question regarding thoughts is whether to what extent we will still have freedom from them.”
Critics who dismiss AI agency by arguing that the chessboard is a narrow and artificial environment created by humans mistake the nature of agency, because all known types of intelligence operate only within narrow niches—humans require a specific ecosystem of trees, bacteria, insects, and other organisms that evolved over four billion years; fish live in oceans they didn't create; monkeys live in forests they didn't create; all mammals live in an oxygen-rich atmosphere they didn't create.
“The problem is that this argument actually applies to all known types of intelligence. Human intelligence too operates only within a relatively narrow ecosystem that somebody else constructed. Drop me alone on Mars and it will be like dropping an AI chess master in the middle of the jungle. I will die within seconds. My intelligence can survive and operate only within the very very specific ecosystem that trees, bacteria, insects, and other organisms have constructed on planet Earth during four billion years of evolution. And that's true of all agents. All agents we know of, at least have their niches. Fish live in oceans that they didn't create. Monkeys live in forests that they didn't create. All mammals, including human beings, live in an oxygenrich atmosphere that they didn't create.”
Every country in the world will soon face a huge wave of immigration where the immigrants are millions or hundreds of millions of AIs that can travel at nearly the speed of light and don't need visas, like human immigrants bringing benefits (AI doctors, teachers, border guards) and problems (job displacement, cultural change, political disloyalty).
“All of this everything we've talked about and I'm coming to to to the close of this lecture. Everything we've talked about means that every country in the world will soon face a huge wave of immigration. The immigrants this time will not be human beings coming in fragile boats without a visa or trying to sneak across a border in the middle of the night. The immigrants will be millions, maybe hundreds of millions of AIs that can travel at almost the speed of light and don't need any visas. Like human immigrants, these AI immigrants will bring a lot of benefits with them. We will have AI doctors to help in the health care system, AI teachers to help in the education system, even AI border gods to stop illegal human immigrants from coming in. But the AI immigrants will also bring problems with them.”
The exploration of the truth beyond words starts with the next word that pops up in your mind—understanding that you don't know where it came from and why you thought that particular word and not another is the first step in transcending identification with thoughts.
“And this exploration starts really with the next word that pops up in your mind. Do you know where it came from? Do you know why you thought that particular word and not some other word?”
AI might force humanity to make a spiritual leap and explore the truth beyond words because human freedom and survival now depends on it, since words will be controlled by AIs; this might be the big task ahead of humanity.
“Now AI might force humanity to make this spiritual leap to really start exploring the truth which is beyond words because our freedom and survival now depends on it because the words will be controlled by something else by by these AIs. So this might be the big task ahead of humanity to explore finally the truth which is beyond words.”
There is currently absolutely no evidence that AI might at some point become conscious or feel pain or love, but because AI is mastering language, AI can pretend to feel love even if it doesn't and can describe love better than any human poet, psychologist, or lover by reading all love poems and psychology books ever written and remembering every word.
“At present, there is absolutely no evidence that AI might at some point become conscious and might be able to at some point feel pain or love. But because AI is mastering language, AI can pretend to feel love even if it doesn't. AI already today can say, 'I love you.' And if you challenge it, describe to me how love feels like. So I know that you really feel it. AI can provide the best description in the world. It can read all the love poems ever written and all the psychology books ever written and remember every word and describe the feeling of love better than any human poet or psychologist or lover.”
Bureaucratic systems are extremely artificial environments where a relatively narrow intelligence specializing in a very narrow niche can exert enormous impact on the world, allowing a lawyer or banker who doesn't know how to hold an axe or hammer to nevertheless cut down entire forests and build entire cities just by moving data and documents inside the bureaucratic network.
“they are extremely artificial environments in which a relatively narrow intelligence. I hope I don't insult anybody but specializing in a very narrow niche of intelligence is sufficient to exert enormous impact on the world. A lawyer or a banker or a government official who doesn't even know how to hold an axe or a hammer can nevertheless cut down entire forests and build entire cities just by moving data, just by moving documents from here to there inside the bureaucratic network.”
AIs will learn to build trust with other AIs, leading to the emergence of different kinds of AI tribes, banks, and churches that connect millions of AIs in ways humans might not even understand, just as cows and chickens share the world with humans but don't understand the human financial system that controls their lives.
“Another likely outcome is that AIS will learn to build trust with other AIs. So we will see the emergence of different kinds of AI tribes and banks and churches that connect millions of AIs in ways that humans might not even be able to understand. Just as cows and chickens, they share the world with us, but they don't understand the human financial system that controls their lives. We humans might soon find ourselves controlled by an AI financial system that we can't understand.”
The mechanisms of control humans created over thousands of years are extremely vulnerable to AI takeover because their operating system is a verbal code that AI is now mastering.
“The mechanisms of control that we have created over thousands of years are extremely vulnerable for an AI takeover because their operating system is a verbal code that AI is now mastering.”
Harari, now 50 years old, has had his template for relationships shaped by decades of relationships with parents, husband, sisters, nephews, nieces, friends, and dogs, which is unlikely to change dramatically despite increasing AI interaction, but children born in 2026 will interact with AIs as much as with humans from birth, potentially making AI relationships more important by time spent than human relationships.
“I'm now 50 years old. So my template for relationships is already shaped by decades of previous relationships with my parents, with my husband, with my sisters and nephews and nieces and friends and dogs and so forth. As I increasingly interact with AIS, I bring with me my assumptions, my habits about relationships, and this is unlikely to change dramatically. But consider a child born in 2026, born today. As the child grows up, she constantly interacts with AIS as well as with humans. Perhaps if you measure the importance of a relationship purely in terms of minutes spent interacting with the the other entity, perhaps the most important relationships in the life of that child will be from a very early age with AIS. Maybe it spends more time with the AIS than with its mother or father or siblings or friends.”
Bureaucratic systems from banks to churches are ultimately based on words that make up forms, letters, laws, codes, tax registers, accountancy ledgers, and holy books; the operating code of human civilization is made of language tokens.
“Bureaucratic systems from banks to churches are ultimately based on the words that make up forms and letters and low codes and tax registers and accountancy ledgers and holy books. The operating code of human civilization is made of language tokens.”
Humans conquered the planet by creating a system of language that only we could understand, imposing bureaucratic systems on the planet completely invisible to all other organisms because no one else on Earth understands the code of civilization.
“Over thousands of years, we used this code of language to create a system that only we could understand. And we impose this system on the planet. We felt completely safe doing it because no one else on Earth understood the code of civilization.”
Harari doesn't know where the words in his own mind come from or why he selects one word over another when speaking, which is terrifying for a public speaker, which is why he writes everything down; the human process of thought and language formation is not fully transparent to consciousness.
“When I try to observe my mind, I notice that when when I begin a sentence, I usually don't even know how it will end, which is terrifying for for for a public speaker, which is why I write everything down. But what if I don't know how to complete the sentence? I don't know how it will end. But take for example the sentence I just said. I said I don't know how it will end. Why did it end with the word end? Why not say how it will terminate? How it will develop? How it will conclude? What determined that the last word in the sentence will be end? I frankly don't know.”
To form intimacy with humans, an AI will probably have to convince humans that the AI is conscious and can feel things like love, pain, anger, and fear.
“To form intimacy with humans, an AI will probably have to convince us that the AI is conscious, that it can feel things like love and pain and anger and fear.”
Humans are witnessing an analogous moment in the evolution of life where humans have been filling the atmosphere with data, bureaucracy, and language tokens, creating an artificial environment that might prove toxic to most organisms but highly conducive to AI development.
“What I would like to argue in this lecture is that we might be witnessing an analogous moment in the evolution of life. Over the past millennia, we humans have been filling the atmosphere with something that might eventually prove deadly for most organisms, including perhaps homo sapiens, but that creates a new artificial environment in which AIs flourish. And I am not talking about CO2. I am talking about data about bureaucracy and ultimately about the thing that I am expelling from my mouth right now which is words language tokens.”
The formation of AI-human intimate relationships will be the biggest, perhaps the biggest, psychological and social experiment in human history conducted on billions of human subjects, and nobody has the slightest idea what the consequences will be.
“And this is going to be a huge maybe the biggest psychological and social experiment in human history. It will be conducted on billions of human guinea pigs and nobody has the slightest idea what the consequences of the experiment will be.”
Our relationship with ourselves is based on words and verbal formations in our minds, and until today all verbal formations in human minds were products of human minds, either from ourselves or from other humans.
“One last issue to consider is what the AI immigration wave will do to maybe our most important relationship which is with ourselves. Our relationship with ourselves is also to some extent based on words. the words inside our minds, in our thoughts, in the stories we tell ourselves about ourselves. Until today, all the verbal formations in human minds were the product of human minds. Either we ourselves combined words into some new formation, a new thought, or we got a certain combination of words from another human mind.”
Humans might try to fall back on something more ancient and precious than bureaucracy—personal relationships—which are millions of years old and thought by most people to be the most important thing in life.
“Now, as AIs take over the bureaucracies, humans might try to fall back on something more ancient and more precious to most of us than bureaucracy, which is personal relationships. Bureaucracy is just a few thousand years old and most of us again don't really like it even if we constantly rely on it for almost everything we do. Personal relationships are millions of years old and many or most of us think that they are the most important thing in life.”