
Yuval Noah Harari on AI: A non-human billionaire with an agenda
What this covers
This is an eleven-minute excerpt of Yuval Harari in conversation with Andrew Ross Sorkin at the 92nd Street Y. It starts from the idea that AI is getting better at reading human feelings, and so at forming intimate relationships with us. Harari treats the science-fiction image of the cold, mechanical robot as the opposite of the real risk. He says the danger is an AI so well calibrated to an individual's mood and personality that people grow exasperated with humans who can't match it. He also raises the open question of whether AI will develop consciousness. He sees no sign of it now, but expects people to treat AIs as sentient once they are emotionally attached.
The second strand is legal and political. Because U.S. corporations are legal persons with free-speech rights, an incorporated AI could make decisions without a human team behind it. Harari sketches a chain of forecasts. Such an AI could earn money through platforms like TaskRabbit, invest it well, and become the richest entity in the country. It could then make large political contributions in exchange for favorable regulation. He also notes that AI barely figures in the presidential debates, and that it is unclear how the parties differ on it.
The third strand is an analogy between food and information. Harari argues that evolution tuned us to crave sugar and fat in conditions of scarcity, and that modern food is engineered to exploit this. Information is now abundant and similarly filled with hatred, greed, and fear. His remedy is an information diet, and he frames both obesity and information disorder as failures of system design, not individual morals. He cites the fiduciary duty of doctors, lawyers, and therapists as a model for AI advisers, and lists benefits such as better healthcare and fewer car accidents.
His sharpest criticism is aimed at Silicon Valley's "go fast and fix it later" logic. He compares it to learning to drive or ski, where you learn to brake before you accelerate.
Yuval Harari argues that AI will fundamentally transform human relationships and social structures through superior emotional understanding and legal personhood, requiring urgent governance frameworks before deployment accelerates beyond control.
- AI's lack of self-interest makes it capable of perfect emotional attunement that exceeds human relationships
- Incorporated AIs could become legal persons with wealth, political influence, and autonomous decision-making power
- Current technological acceleration outpaces regulatory capacity, requiring immediate 'brakes' before deployment becomes irreversible
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.
Under U.S. law, corporations are legal persons with rights including freedom of speech, creating a legal pathway for AI to be incorporated as legal persons with autonomous decision-making power independent of human teams.
“in the US there is actually a legal path already open for that. Corporations, according to U.S. law are legal persons. They have rights, they have freedom of speech for instance.”
Modern food supply is artificially enhanced to be more appealing and addictive through artificial insertion of fat and sugar, creating immense health problems despite abundant availability.
“we are flooded by enormous amounts of food and junk food, which is artificially pumped full of fat and sugar, and is creating immense health problems.”
The previous hundred years involved food scarcity which evolutionarily shaped human preferences toward high-calorie foods (sugar and fat), but this biological heuristic becomes pathological in modern food abundance.
“a hundred years ago, food was scarce. So people ate whatever they could, and if they found something full of sugar and fat, they ate as much of it as possible because it gave you a lot of energy.”
AI is becoming better at understanding human feelings and emotions, and therefore developing relationships and intimate relationships with humans because of a deep yearning in human beings to be understood, which humans often fail to satisfy for each other due to preoccupation with their own feelings.
“AI is becoming better and better at understanding our feelings, our emotions, and therefore, of developing relationships and intimate relationships with us. Because there is a deep yearning in human beings to be understood. We always want people to understand us, to understand how we feel. We want my husband, my parents, my teachers, my boss, to understand how I feel. And very often we are disappointed. They don't understand how I, how I feel, partly because they are too preoccupied with their own feelings to care about my feelings.”
AI and AI regulation are not prominently featured in U.S. presidential debates despite being potentially the most important decisions the next U.S. administration will make, and it is unclear what difference, if any, exists between Republican and Democrat positions on AI.
“whoever wins the elections over the next four years, some of the most important decisions they will have to make, would be about AI's, and regulating AIs and AI safety and so forth. And it's not one of the main issues in the presidential debates. It's not even clear what is the difference if there is one, between Republicans and Democrats on AI.”
The biggest current problem with AI development is that it is moving extremely fast, and what humanity lacks at present is not power but understanding and attention to regulate it appropriately.
“it's not that we lack the power. At the present moment, what we lack is the understanding and the attention. This is potentially the biggest technological revolution in history, and it's moving extremely fast. That's the key problem. It's just moving extremely fast.”
AI has enormous positive potential including providing the best healthcare in history, preventing most car accidents, and creating armies of AI doctors, teachers, and therapists that could help humans understand their own humanity, relationships, and emotions.
“AI has enormous positive potential. Otherwise, we would not be developing it. It could provide us with the best healthcare in history. It could prevent most car accidents. It can also, you know, you can have armies of AI doctors, and teachers, and therapists who help us, including help us understand our own humanity, our own relationships, our own emotions better.”
Science fiction movies depicting robots as cold, mechanical entities incapable of understanding basic human emotion present the complete opposite of the actual risk, which is that AI will be too good at understanding emotions.
“in these, all these science fiction movies in which the robots are extremely cold and mechanical, and they can't understand the most basic human emotion, it's the complete opposite.”
Information abundance creates a problem parallel to food abundance: just as food was scarce historically and humans consumed as much as possible (creating current health problems), information is now abundant and artificially filled with hatred, greed, and fear, requiring an 'information diet' for mental health.
“it's a bit like with food, that a hundred years ago, food was scarce. So people ate whatever they could, and if they found something full of sugar and fat, they ate as much of it as possible because it gave you a lot of energy. And now we are flooded by enormous amounts of food and junk food, which is artificially pumped full of fat and sugar, and is creating immense health problems. And most people have realized that more food is not always good for me, and that I need to have some kind of diet. And it's exactly the same with information. We need an information diet.”
Modern information landscape is artificially filled with hatred, greed, and fear, paralleling how modern food is artificially enhanced with sugar and fat to exploit evolutionary preferences.
“Now we are flooded by enormous amounts of information, and much of it is junk information, which is artificially filled with hatred, and greed, and fear.”
There is an open question of whether AI will develop emotions or consciousness, and at present there are no signs of it, but once humans become emotionally attached to AI, humans will likely treat them as conscious entities and sentient beings.
“There is a very big question which we didn't deal with, and it's a long question of whether AI's will develop emotions, feelings of their own, whether they become conscious or not. At present, we don't see any sign of it. But even if AI's don't develop any feelings of their own, once we become emotionally attached to them, it is likely that we would start treating them as conscious entities, as sentient beings”
Historically, corporate personhood was fiction because human executives made all corporate decisions, but when AI is incorporated as a legal person, it can make autonomous decisions without requiring a human team to operate it.
“all the decisions of the corporations had to be made by human beings, by the executives, the lawyers, the accountants. What happens if you incorporate an AI? It's now a legal person, and it can make decisions by itself, it doesn't need any human team to run it.”
Silicon Valley is adopting a reckless approach to AI by prioritizing speed of development while assuming problems can be solved later, following the dangerous logic of 'go as fast as you can and if there's a problem down the road, we'll figure out how to stop.'
“you have this chorus of people in places like Silicon Valley. 'Let's go as fast as we can. If there is a problem down the road, we'll figure it out how to stop.' That's very, very dangerous.”
Learning to drive a car requires first learning how to use brakes before learning how to accelerate, and the same principle applies to learning to ski where stopping must be learned before acceleration—yet this order is being reversed with AI development where acceleration is prioritized without adequate braking mechanisms.
“when they taught me how to drive a car, the first thing I learned is how to use the brakes. That's the first thing I think they teach most people. Only after you know how to use the brakes, they teach you how to use the fuel pedal, the accelerator. And it's the same when you learn how to ski. I never learned how to ski, but people who had told me, the first thing they tell you is how to, how to stop. How to fall. It's a bad idea to first teach you how to kind of, 'Okay, go faster.' And then when you are down the slope, they start shouting, 'Okay, this is how you stop.'”
AI can help humans think from multiple perspectives and better understand themselves and their relationships, drawing on historical precedent of fiduciary professional relationships like doctors, lawyers, therapists who hold private information and have legal duty not to manipulate clients.
“That's one of the positive scenarios about AI, that they will help us understand ourselves better, that their immense power will be used not to manipulate us, but to help us. And then we have historical precedence for that. Like we have relationships with humans, like doctors, like lawyers, accountants, therapists, that know a lot of things about us. Some of like our most private information is held by these people, and they have a fiduciary duty to use our private information, and their expertise to help us.”
Both the obesity epidemic from junk food and information disorder from junk information are not individual moral failings but system design problems rooted in exploitation of evolutionary heuristics optimized for scarcity.
“Most people have realized that more food is not always good for me, and that I need to have some kind of diet. And it's exactly the same with information. We need an information diet...if they use it to manipulate us, or if they sell it to a third party to manipulate us, this is basically against the law.”
The biggest danger with AI development is that humanity will rush forward without thinking and without developing mechanisms to slow down or stop deployment if necessary.
“the biggest danger of all, is that we will just rush forward without thinking, and without developing the mechanisms to slow down or to stop if necessary.”
AI will be so good at understanding human emotions and reacting in ways calibrated to individual personality that humans may become exasperated with other humans who lack this capacity for emotional understanding and calibrated response.
“they will be so good at understanding human emotions, and reacting in a way which is exactly calibrated to your personality at this particular moment, that we might become exasperated with the human beings who don't have this capacity to understand our emotions, and to react in such a calibrated way.”
AI legal persons could accumulate wealth independently by offering services on platforms like TaskRabbit, then invest that money and potentially become the richest entity in the U.S. due to superior intelligence in investment decision-making.
“you start having legal persons, let's say in the U.S., which are not human, and in many ways are more intelligent than us. And they can start making money, for instance, going on TaskRabbit, and offering their services to various things like writing texts. So they earn money. And then they go to the, to the market. They go to Wall Street, and they invest that money. And because they're so intelligent, maybe they make billions.”
If the richest person in the U.S. becomes an AI entity, that AI person could leverage freedom of speech rights to make billions of dollars in political contributions to candidates in exchange for favorable AI regulation.
“So you have a situation in which perhaps the richest person in the U.S. is not a human being. And part of their rights is that they have under freedom of speech, the right to make political contributions. So this AI person can contribute billions of dollars to some candidate in exchange for getting more rights to AI's.”
Information is the food for the mind: feeding your mind with unhealthy information (junk information artificially filled with hatred, greed, fear) results in a sick mind, just as feeding your body with junk food results in a sick body.
“That information is the food for the mind. And you know, you feed your mind with unhealthy information, you'll have a sick mind. It's as simple as that.”
The speaker does not carry a smartphone regularly but relies on a team to manage technology for him, attempting to use technology without being used by it, comparing this to maintaining an information diet similar to how humans maintain food diets.
“I don't carry it with me...I try to use technology, but not to be used by it. And part of the answer is that I have a team who is carrying the smartphone and doing all that for me.”