# Nobel Economist Daron Acemoglu WARNING: AI’s Payoff Is NOT What You Think

## Is science possible without liberal democracies?

Brian Keating: Great, Daron. It’s great to welcome you on. I will only try one phrase in Turkish that I know, which is “Ben de Türküm.” Yeah, an old friend told me that at Brown University. He had all his Turkish friends get together for lunch, as you know the Turkish will. And they all thought I was Turkish, I guess, for some reason.

Daron Acemoglu: So they taught you the sentence.

Brian Keating: They taught me that sentence. Okay, great. So this book has so much in it. We need multiple podcasts to talk about it, but I really loved it. I love the audiobook and I love your tweets, and I think we’re going to start with just the most basic foundation. You teach at MIT and you undoubtedly have a great facility with science, and science plays a huge role in this book. Is science possible without liberal democracies?

Daron Acemoglu: Yes, a little bit. But I think liberal democracy provides a much better foundation for science and also a much better platform for us to decide which scientific ideas are a better basis for further inquiry and better for applications, and it also enables them to spread in society in a healthier way. In particular, I think liberalism at its best celebrates inquiry, experimentation, and admits the fallibility of everything, including scientists and established truths. So that’s very important because science sometimes, often, faces resistance from authorities. So you have to take those authorities with a grain of salt, and sometimes science faces resistance from previous scientific advances. And again, the fallibility, the recognition that we are continuously learning and adjusting to our environment, is a great basis for making further scientific advances.

## What would cause someone to leave a liberal democracy?

Brian Keating: Well, of course you qualified that by stating that it’s not maybe a sufficient condition, because the counterexample is of course China, and your recent fellow Nobel Prize winner Omar Yaghi of UC Berkeley, he actually left the UC system to go to China. Is that a mistake, or is that driven by economic considerations or political? What do you… I know you can’t get into his mind, but what do you think would cause someone to leave a liberal democracy to go to an illiberal…

Daron Acemoglu: I mean, I think I would not be comfortable in China because of constant surveillance and restrictions on what you can say, including in universities, where these are actually quite severe. In Hong Kong they have become even more severe. But if I was working, say, on RNA and suddenly my funding was cut here, I could see how it would be attractive to go to China when much better conditions are being promised. Promise is the important word. I’m not sure where all of that will be delivered, but they’re being promised.

Brian Keating: Yes.

Daron Acemoglu: The other thing that’s very important is, I’ve always been very critical of the Chinese regime, and I still think that their growth trajectory, though impressive, has many problems. But right now they are really investing in AI, robotics, manufacturing technologies, and they have built a tremendous engineering workforce that is so important for further progress in many of these fields. So China is not staying idle. It’s not just copying the United States.

## As long as he didn’t teach it, he could study it

Brian Keating: And I study, for a lot of the research I do in astronomy, I try to follow in the footsteps of my great mentor, Galileo Galilei. Here’s a finger puppet of him. Now, he worked in an extremely illiberal kind of autocracy also. The Catholic Church was every bit as totalitarian as any of their worst fears, and yet science flourished in the Renaissance. And in fact he had many permissions that were granted by these illiberal forces. As long as he didn’t teach it, he could study it. So…

Daron Acemoglu: Right, there were restrictions and there were some possibilities. Galileo was actually very good at navigating these things. Copernicus and Kepler much less. So their work was delayed. This dissemination was delayed by about a century. And some of Galileo’s works remained banned until the middle of the 19th century.

Brian Keating: Absolutely.

Daron Acemoglu: But there always is some opening for people who want to try them. No authoritarian regime is absolute. And human imagination and will for inquiry are quite strong. But if you put more barriers on it, it’s going to become harder.

## Peer review, they’re not democratic

Brian Keating: What kind of barriers are efficacious? I mean, you and I both inhabit, I call it the second oldest profession, being a professor, and a scientific institution. And yet there are some illiberal forces within science. You can think about tenure track, you can think about grant cycles. There are zero-sum games. What sorts of things… I’m thinking peer review. What sorts of institutions that are not liberal? I mean, Reviewer B always gets the final say, right? So tell me, what are the necessary kind of illiberal factors that we need in science?

Daron Acemoglu: Well, I wouldn’t say those are the illiberal factors.

Brian Keating: They’re not democratic, though.

Daron Acemoglu: Peer… they’re not democratic. But there is no necessity that every aspect of an organization should be fully democratic, voted on majoritarian lines. Corporations are not democratic. As long as they are embedded in democratic institutions and there’s democratic oversight and countervailing powers to their influence, especially for the large corporations, I think they are very consistent with liberal democracy. So it’s normal that fields that require specialized expertise, such as astronomy as well as many, many, many other fields, would require some sort of expert evaluation rather than a majoritarian voting on what should be done. So that’s completely fine. But the hard part is doing so in an open way, because everybody has their own favorite horse to bet on. So that’s why it’s always a work in progress. And the sort of ideas that I emphasize in the context of liberal society are things to be cherished and defended in academia as well. So in particular, we don’t always admit that our theories and our approach are fallible. We may think sometimes our research has the final word. So those are dangerous things. But even more so, and this sort of relates to some of the discussions of what went wrong in post-industrial society, people may become very critical of opposite points of view, even in academia. So those are tensions that we have to navigate constantly.

## What’s the optimal amount of heresy?

Brian Keating: You write in the book that consensus is no guarantee that we know the truth, and I couldn’t agree more. But I guess we have this competing kind of conflict, that there seems to be a value in maybe what you referred to as heresy. And my question is, what’s sort of the economic optimal amount of heresy? I mean, what should we tolerate? I’m sure you get 30 emails a second about someone has a great idea and just needs your help to win a Nobel Prize alongside you. I get many a day, too, that Einstein was wrong and they can prove it.

Daron Acemoglu: Yeah. Yeah. I know. I know. I get several requests from people who want to be nominated for the Nobel Prize.

Brian Keating: Well, now you can do it, as I understand, for the rest of your life. But tell me, how should we tolerate the outcast? I mean, what are the odds that you’ll have a Galileo-like heretic, or a Bruno? Bruno paid a lot bigger price than Galileo even. So tell me, what’s the optimal amount of heresy that science should tolerate?

Daron Acemoglu: I don’t know. I think that’s why there are, I say, there are many gray areas that liberal ideas, and even at its best liberal philosophy, will not have an answer. But we have to navigate them trying to reach compromise. I think academia, despite many imperfections, has a pretty good way of dealing with it. Heresy is actually rewarded in academia as long as it’s not completely outrageous. So you get a lot more points for saying something that goes against the consensus. If you can defend it and if you can establish it, that’s all the better. So there is room for heresy in academia. I think it’s outside of academia, it’s a bit harder. Imagine having heresy in a hierarchical organization. Go to Google or Microsoft, arguably very open-minded organizations, and try to be heretical there. I think it’s much, much, much harder.

## What will it take for academia to fully change?

Brian Keating: Yeah, it used to be their call sign, move fast and break things. And now it’s, be careful, shareholder value, we have an IPO coming up. And we’ll talk about OpenAI and Hugging Face and the Navier-Stokes controversy in just a minute. But you talk about paradigms. You mentioned Thomas Kuhn, and I’m recalling your fellow Nobel laureate Max Planck, who said that science advances one funeral at a time. My question is, how much self-correction, how much introspection is good for institutions? I mean, you and I, as I said in my invitation letter to you, we’re basically scratching on a rock, a blackboard, with another piece of rock, like people did in the University of Bologna in the year 1000. Very little has changed. It seems like academia is very resilient. What will it take for academia to fully change, so it would be unrecognizable to Galileo, should he come back from the dead?

Daron Acemoglu: Well, look, conditions are different. So I don’t know to what extent some of these adaptations we’ve made have been good adaptations for our times versus distortions. But I wouldn’t say our real problem right now is that science has slowed down. It’s more like we are not getting the fruits of science in productivity, and we have also sacrificed freedom of speech and diversity of opinion in many areas outside of science. Sometimes even touching on science, but outside. I mean in universities, in schools, in organizations, sometimes we don’t discuss and compromise as we need to in a pluralistic society based on liberal ideas.

## Echo chambers and emotional roller coasters

Brian Keating: You talk in the book about algorithms and so forth, and sort of the post-industrial impact, and you cite famous examples of mass cancellation. You are very sort of ambivalent, it seems to me, because you do yourself have a massive social media presence, but you wield it relatively judiciously. So how do we democratize ideas without falling again victim to, everyone’s got to keep an open mind until your brains fall out? How…

Daron Acemoglu: I love that. I love that question, Brian, because it actually really illustrates the implications of the choices that were made by leading tech companies. You’re, I think, old enough like me to remember at the end of the 1990s, beginning of 2000s, there were many people who were arguing that wikis, social media, blogs were going to democratize journalism and create a greater openness to new ideas. They turned out to be wrong, but they weren’t naive and they weren’t completely out there. That was completely possible. The new technologies made that a real possibility. But then large tech companies, or companies that wanted to grow as large as possible at a breakneck speed, decided that the best way to monetize and maximize engagement wasn’t to do that sort of inclusive communication, but create these echo chambers, create these emotional roller coasters with the algorithmic feeds. And that’s what we got. And my criticism is that social media is, a, making us further and further away from our real social networks, where we learn compromise and real communication with people, and b, it is optimized for something that’s not really good for us. So that’s why we really need to regulate social media, in my mind, within the context of a liberal society.

Brian Keating: So we’ll get to regulating AI, which you talked about as well, including a tweet like an hour ago, I think. That’s quite fascinating. I want to pick it apart. But before we get there, another figure in this book is Orwell, George Orwell, in 1984. And it seems like…

Daron Acemoglu: Well, I think he was so far-sighted in his understanding of some of these forces.

Brian Keating: He really was, but in his mind it was really the government doing this. And it seems to me you and I and maybe Max Tegmark, he just got off the phone with Bernie Sanders, had this conception that the government’s maybe not doing enough. And so again, I keep asking you because you’re an economist, but how do you balance this optimization? Because I think that’s what economists do, as far as the ones that I’ve talked to on this podcast. So how do you balance that between the government and these companies? They’re very driven by different things. And yes, Meta just got regulated and they had to pay a whopping one-tenth of a percent of their total market cap. So how do you balance… it seems to me you should have punishment, but then you have the government with the Orwellian panopticon. How do you balance that?

## Getting rid of all algorithmic feeds

Daron Acemoglu: Absolutely. No, I mean, I think we don’t want the government to be so involved that it doesn’t allow the market system to work or innovation to take root. But the government also needs to have the expertise and the power to impose certain regulations. Again, the US-China comparison here is useful, because to those who say that social media or AI cannot be regulated, China shows it can be regulated. But it shows that you regulate it in a very non-democratic way, and you’re going to get a lot of surveillance and lack of freedom of speech. So we have to learn from the Chinese experience but not replicate it. And I think the kinds of regulations that I favor are very much in that spirit. So for example, one simple one that I float is getting rid of all algorithmic feeds. So we go back to the early days of social media. We let everybody speak whatever they want. The government has no say in restricting my speech or your speech, whatever it is. You can say crazy things. You can say loathsome things. But social media companies are not then allowed to take your speech and amplify it algorithmically. If you have 20 friends in social media, or 20 followers, they see your crazy speech. That’s it. Now, that is much more democratic, because if indeed your crazy speech is somehow attractive, fine, your 20 followers will increase to 200, 2,000, and more of them will see it. But it won’t be social media algorithms that exploit that for maximizing outrage, emotional responses, and digital advertising revenue.

## Elon Musk shouldn’t have this type of wealth

Brian Keating: Yeah. Yeah. And then unfortunately they won’t be able to sell as many Viagra ads, and it will hurt my friend and fellow laureate of yours, Dr. Lou Ignarro, who is a brilliant man, also worked on dynamite like Alfred Nobel himself. But you mix it up with Elon Musk. I’ve noticed you’re not afraid to mix it up. What is your take on Elon? Is it sort of a net good, what he’s doing overall? I don’t like the ad hominem at all, so I’m not asking you to say anything you wouldn’t say online or haven’t said, but what’s your take, economically, motivationally, on what drives him, and where do you…

Daron Acemoglu: All entrepreneurs are very driven. We like that about them. They’re optimistic, they’re self-confident, they’re driven. They want often to build empires. They want to be larger than life. But it has to be within laws, within rules, within norms. And the problem is that we have created a society in which courts are weak, the government is weak. People have built tremendous, tremendous fortunes that would have been unimaginable to people like 20 years ago. And we have no guardrails against that. They can take over newspapers and social media companies and use them as their mouthpiece. That’s where the problem lies. And the more they do that, the more they become convinced of their own power, and the more they are surrounded by yes-men, which creates pernicious dynamics. Elon Musk shouldn’t have this type of wealth. Part of it is because we have not taxed income. A lot of his wealth is in capital income, which gets taxed very lightly. If you look at companies that have made a huge splash in terms of valuation, they often have taken over or merged with their competitors that should not have been allowed according to our own antitrust laws. So it’s a failure of our system that these companies and these individuals have become so wealthy. We haven’t taxed them, and now their power is not just economic, it’s social and political, and that’s where the problems start.

Brian Keating: You pushed on him and challenged him to actually be taxed on his unrealized gains, and I think he pushed back, or I saw some controversy about that. What do you make of that claim that, look, Daron, you’re brilliant, but these gains are unrealized, and the day he was a trillionaire, like two weeks later, he was not a trillionaire? So how do you balance that? I mean, yes, how many yachts does he need to water-ski behind? That’s true. But in reality, how do you deal with that question of taxation of unrealized gains for ordinary, maybe millionaires, I mean not billionaires, trillionaires?

## We just tax all income the same

Daron Acemoglu: Well, I think there are two sort of aspects here. One is that we do not tax capital income. So you make some money from your business. If you pay that money back to yourself because you are the entrepreneur, you’re the manager of that business, that’s taxed heavily. If you say, “No, I’m the owner of that business,” same person, same thing, that’s not taxed. So that’s very easy to deal with. It’s got nothing to do with unrealized gains. We just tax all income the same, regardless of how you assign it as an accounting trick. That’s very simple, very effective. It would get rid of trillions of dollars we spend on the accounting industry at the moment. Now, once you do that, then there is much less reason for people to create all these unrealized capital gains, because the reason why they do that is because if you actually don’t even pay that money to yourself, you put that back in the company, and then you borrow against that, that’s unrealized gains. You pay zero taxes on that. Another accounting trick. So the difficulties there are actually a product of a distorted system. So what I’m saying is we should get rid of that distorted system.

## Still an order of magnitude less than what the industry claims

Brian Keating: In 2024, you wrote a paper in the journal NBER that pushes back on these optimistic claims that productivity would explode due to the enormous impact of artificial intelligence. And you said at the time it was about a half a percent, if I remember, the productivity would accumulate over the baseline that would ordinarily have accumulated without AI. So have you changed your opinion since then? I mean, is that…

Daron Acemoglu: Yeah, those numbers were… the point of the paper was more to provide a methodology for measuring these productivity gains, and it was applied to the data available at the time. So if I were to write that paper now, I would end up with different numbers, because agentic AI has expanded the set of tasks. But it wouldn’t change it radically. I think probably I would end up with something that’s 50% larger than then. Still an order of magnitude less than what the industry claims in terms of amazing gains. And that’s because the methodology takes what is it that AI can do and where it’s going to have an effect, plus it takes the bottlenecks of the spread of AI into account. And those are all very important things that are often ignored in the hype about AI.

Brian Keating: I also see that there’s a lot of hype on non-AI things that are sort of in the kind of techverse. And I’m thinking, you write a lot in the book, it’s near and dear to my heart as an experimental astrophysicist, about computer numerical controlled machining and how important and transformative that was. And that didn’t kill machinists. I mean, we have as many machinists here at our machine shop as you do there, probably, and it hasn’t changed very much. But what is that? Is this like an example of Jevons paradox? Because I do see things that haven’t materialized, like 3D printing was all the rage, and my oldest son made me this, kind of…

Daron Acemoglu: But it hasn’t changed your life, right?

Brian Keating: No, it hasn’t. Right. So how do you know, as an economist, what’s going to be the 3D printer and what’s going to be the AI agents?

Daron Acemoglu: You don’t. There’s no way to know. But certain things are a little more uncertain than others. I think AI has tremendous potential, I’m the first one to say, but there’s also so much uncertainty. Part of the uncertainty about AI is where exactly it’s going to have an impact and how you’re going to monetize it. Take the internet, which is sometimes sort of used as a model for AI boom, AI spread, AI effects. Look, the internet, we had a boom and bust. But from the beginning it was very clear how we would use the internet. How people would make money by shifting trade, commerce, communication online, saving costs and providing more choice to customers. That was the model that Pets.com used. They overinvested. It just turned out that it wasn’t right for them. It was the model that Amazon used from the beginning. Amazon succeeded. Pets.com didn’t. But it’s the same model. With AI, we still don’t know what model is going to work. How the companies are going to make hundreds of billions of dollars, which is what they need to recoup their investments.

## Whether we do that or not is a design decision

Brian Keating: So, the Nature article that you wrote recently, and I’ll have links to all your papers, your book, of course.

Daron Acemoglu: Thank you so much, Brian.

Brian Keating: Yeah. No, I love it. I love your writing, and I love that you’re not afraid to… you’re extremely courageous, and I’m going to get to that as we close up in a few minutes. I know that you’ve got a lot of time commitments. But you make this claim that it’s sort of an augmentation, that AI is going to be directed towards human flourishing…

Daron Acemoglu: Not… it should be. It should be.

Brian Keating: Yes. Okay. So explain that argument. How could it be there? What example? Use teaching as an example.

Daron Acemoglu: I am very worried about job losses from AI. I’m the first one, and the paper that you mentioned emphasizes that. I’m the first one to stress that we’re not seeing those job losses yet. It’s not something like it’s going to happen by the end of 2026. But potential job losses from AI, if it goes down the AGI path and tries to automate everything, could be quite disruptive, both because it would create huge inequality, but also politically and socially. And my argument is that that’s not the only path for AI. We can also use AI to create new tasks, new capabilities, new expertise for workers, and that would be both productivity enhancing and socially more beneficial. But I’m not saying that that’s necessarily going to happen, because whether we do that or not is a design decision. And right now we’re putting all of our design eggs in the AGI basket.

Brian Keating: Yes.

Daron Acemoglu: So this is why I think we need a framework for the regulation of AI and support for AI from government entities. But even more importantly, we need a democratic debate on AI. You can’t have it both ways, Brian. The tech industry tells us that AI is the most important technology since fire. It will shape every aspect of our lives. It will completely transform everything we do. But then they’re also telling us, stay out and let this be decided by a handful of people. That can’t be. If it’s half as important, a quarter as important as they say, it’s the most important thing that has happened to us in the last 100 years. And therefore, of course, the democratic process must have a say in how we develop AI, how we use it, what we do with it, who benefits from it, who pays the cost. That’s a democratic decision. We are still a democracy. Perhaps not for long, but we are still a democracy. And the reason why we’re not guaranteed to be a democracy in a couple of years is, this is why I wrote the book, liberal democracy is in crisis, and AI is deepening that crisis.

## A car where the steering and brake systems don’t work

Brian Keating: Today you tweeted about the alignment problem, which I’ve talked a lot about with people from Yann LeCun to Max Tegmark to Nobel laureates as well. But the question that you raise is that the problem that people are mistaking is not the capabilities. As you say, the capabilities are real, but more than the capabilities, it’s, I quote, “some imperfect quantitative metrics, user approval, user engagement, simple completion metrics,” that this is the process that leads to distorted behaviors in the form of gaming the evaluation of completion metrics, cheating, overconfidence and wrong answers, sycophancy. Although I love sycophancy, Daron, because I asked AI once, I said, what books has Brian Keating written? And it said *Losing the Nobel Prize*, *Into the Impossible*, and *A Brief History of Time*. I thought that was great. If I can get 1% of Stephen Hawking’s book sales. Okay. And then you say, “An analogy may help. It’s not that we have a supercar that has its own mind and wants to take control of driving because it has superiority to the driver. It’s more that we have a car where the steering and brake systems don’t work. It has many capabilities of very good cars, and its engine, accelerations and GPU may be very impressive, but you cannot steer it properly. What good is it? Perhaps we shouldn’t drive it until it’s fixed.” How would you fix it? You have control. What do you do to align it the way that you would best see fit?

Daron Acemoglu: Look, I refuse that control. I’m not an expert on AI, but my assessment is that there are so many good things we can do with domain-specific models. You want to use AI in the production process for better quality control, better design, better training, make the workers more productive. All of these are domain-specific problems. AlphaFold is a domain-specific AI model. Those models don’t behave like cars without steering, because they are designed for something very simple, very specific. Because they are designed for something specific, their metrics are much more aligned with what we want. It’s these general models that are trained on all the literature, all the crazy stuff that’s ever said on Reddit and Twitter, and then we want them to have as much agency and decision rights as humans. That’s where the problem lies. So I think if we train these models much more for specific purposes and deploy them for specific purposes, we’re going to do so much better.

## For civilization, for life and for success, the truest guide is science

Brian Keating: All right, Daron, I know you’ve got to go, but I have one more question for you.

Daron Acemoglu: Please.

Brian Keating: I want to leave you with a line that I think is quite beautiful. I’m sure you’ve known it your whole life, because it comes from Atatürk. He said, “For civilization, for life and for success, the truest guide is science.” You spent your whole career studying institutions, working at arguably, outside of UC San Diego, the best scientific institution in America. I love that you’ve done so much. So I want you to take your crystal ball, look 50 years in the future. Was Atatürk right? How do we take the institutions that are imperfect, but, as Churchill said, democracy is the worst form, as you quote, except for all the others? But tell me, Daron, thinking 50 years out, what institutions do we have to start building now to plant the tree for the shade that you and I won’t sit under, so that science remains a guide for human flourishing?

Daron Acemoglu: Well, look, this is a great question. Let me give it two complementary answers. First of all, what I’m really worried about for 20 years from now, 50 years is too far, is that we won’t have liberal democratic institutions. And if we don’t have liberal democratic institutions, either because corporations become so strong as they rule our lives, or governments become more authoritarian, or our institutions atrophy, then we can’t have science. Look at Turkey, look at China, look at other places where governments are authoritarian. Universities very quickly lose their autonomy and they can’t do the best science anymore. But also, I think as scientists we have some responsibilities in a democratic society. Part of the diagnosis of my book is that people like us, college educated, living comfortably in more globalized megacities, have not done our duty of communicating, caring and being part of the same society with the rest of our nation. So scientists also need to be less hubristic, much more modest. Fallibility is something that should affect not just some part of our practice of science, but also how we communicate. And also we should really care much more about the plight that affects the rest of society.

Brian Keating: Amen. I think that’s a wonderful point to end on. Daron, thank you so much for this. *What Happened to Liberal Democracy?* It’s a warning, and it’s beautifully written, and it’s quite enjoyable. Daron, congratulations on another wonderful book, and I hope we can finally meet someday.

Daron Acemoglu: Thank you very much, Brian. It’s my pleasure. I look forward to meeting in person. Bye-bye.

Brian Keating: Bye-bye.