In this talk Tyler Cowen argues that the advent of strong AI will not eliminate jobs but will radically reshape them. He outlines a dozen sectors where new employment will surge—energy infrastructure, data curation, biomedical and other experimental work, entertainment (with a premium on human charisma), artisanal construction, niche service roles, elderly care, and a wave of ultra‑small AI‑enabled startups. Cowen stresses that the speed and direction of these changes depend on a country’s labor‑mobility and retraining capacity; the UK, he notes, faces high energy costs and a need to mobilize talent across regions. He warns that many traditional high‑skill, “conscientious” career paths will see their relative advantage erode unless workers become “AI natives.” A new class of “AI maniacs”—people who master AI orchestration rather than model building—will capture outsized wealth. Finally, he highlights a short‑term paradox: while AI promises massive leisure gains in the long run, the next five years will demand unprecedented effort to learn and integrate AI tools. The talk ends with a call for rapid policy action on mobility, retraining, and regulatory frameworks to capture these opportunities.
Chapters
The era of strong AI is here, but you're all going to have jobs. I suppose that's the good news and the bad news. If you think about earlier English history before the Industrial Revolution, about 80% of the population is in agriculture. And if you ask people back then, what are your new jobs going to be? I don't think one of them would have said podcaster or even person standing up in front of this room giving a talk necessarily. So jobs evolve. Let me talk about a few areas where there will be a lot of new jobs in an AI-drenched world, and I'll talk about how they apply to the UK also. First, energy. There will be a lot of jobs in the energy sector. Every major country needs to build out its energy infrastructure. If you want job growth, a person who, say, right now might become a consulting partner for McKinsey may in the future become an energy company executive and live in Houston rather than New York City. The energy sector is one that worries me the most about the United Kingdom. You probably all know about the plans of OpenAI to put a major data center in the UK. It was ascertained that the cost of energy here is about 4 times higher than in Texas, and as of now, the plan is called off. So I would say if there are jobs problems in a world of strong AI, they are self-inflicted. But a country needs to invest in its own labor mobility across regions, across sectors, across kinds of jobs. If you want one policy recommendation, It is that countries that have low labor mobility will end up in bad shape because a lot of jobs need to change or be reallocated. Countries that have relatively high labor mobility will thrive and flourish. So many jobs— the job itself, the title won't change, but what you do will change. So, so many jobs involve moving in the physical world. You go somewhere, you go to a meeting, You persuade a colleague, you pat someone on the back, you take someone aside, you travel to meet a client. Those jobs, for the most part, will not go away. But in all of those jobs, virtually all of those jobs, the people doing them will need to learn how to use AI. This is the single biggest adjustment of this new era. Everyone, really everyone, will need to retrain themselves. So there's no reason not to start undoing this as quickly as possible. And countries that are very good at retraining will boost their labor productivity rather than having a lot of confused, alienated workers. Another big new growth area for jobs will just be gathering data. So the major AI models, yes, they've read the internet, quality adjusted, but there's so much data out there in governments, in universities, in companies, in archives that's not fed to the AIs. Just putting that data in a form that is readable and feeding it to the AIs again will be a very large sector in the pretty near future. AIs love data. Most of the world's data is just sitting somewhere being wasted. So this is a good example of a job you may not think of readily, but again, a major growth growth area. Uh, UK is actually pretty well situated to do that. You have incredible universities, all sorts of data, plenty of excellent companies, business services. Uh, in theory at least, you could be world leaders in that area if you show the initiative and the mobility. Also, your health systems, uh, however much people may complain about them, when it comes to gathering data, They are top tier for the world. So again, this is a major area where the UK could see a lot of growth. Just running experiments will be a massive growth sector. So AI systems will have, or already have, so many new ideas for medical devices, for drugs, for procedures, but they're not tested. Some of them maybe you can test by simulation, but for the most part, you need to run experiments. So I think the demand to run experiments of all kinds, but especially biomedical, will go up, not by 20%, not by 30%, but maybe by 10x. So what will be a major economic sector simply running experiments? So do you have a regulatory regime that allows your private sector to take, say, UK-generated healthcare data and then actually run experiments on different things During the pandemic, UK did great on this, so I am hopeful. But again, for the future, it is always an open question. There will be many, many new jobs running experiments. Just entertainment will be another growth area. Eventually, people will have more leisure time. They'll want to have more fun. There'll be more disposable income. Now, you might say, well, won't the music be composed by AI, or won't the painting be generated by AI? I'm not sure. I think humans have a desire to consume the emotionally driven outputs of other human beings. Like, I don't want an AI Beatles or AI jokes that are as funny as Monty Python. They're not actually as funny to me. But even if we're content with the AI outputs, if there's some, say, Taylor Swift or Oasis of the future, there will still be human front people. So humans with charisma, humans who are attractive, humans who are memorable will be front people for the creations of AI or the joint creations of AI and humans. So just ask the question, well, how charismatic are the British right now? I leave this as an exercise for all of you. But you have a pretty good history in this regard. James Bond, the Beatles, I mentioned Monty Python, Oasis, so on and so on. Maybe there's been a slight fall-off in that stream as of late, but I'm confident you can all get it back. But just huge returns to charisma in an AI-strong world. Artisanal construction, just things that are made uniquely by craftspeople who have some talent that is special to them. I think that will grow relative to mass production. People will want things that are special, interesting, original. That they wouldn't have seen coming out of what we sometimes call AI slop. Another growth area for jobs. Just weird service jobs. People who are nice to you, your personal assistants, greeters at the restaurant. Uh, there's always room for more weird service jobs, especially as we're gonna have more and more people who are very wealthy, and very wealthy people tend to have a kind of retinue of people who help them do different things and Some of that will be AI agents, but I think a lot of it will be run by people with weird service jobs. In general, it's the parts of jobs that are tedious and repetitive that will go away and be automated first. That's a longstanding trend. So most jobs will become more original, more unique, harder to describe. The harder to describe part is one reason why people are pessimistic. Uh, factory line assembly, more and more will be done by robots and AI. That's a job that's easy to describe. A job like mine, I do 7 different things in the course of a day, one of which is go around giving talks. You could have had ChatGPT give this talk for me. It knows Tyler. You could have just set it up. Don't fly me over. ChatGPT, what would Tyler Cowen say about AI and jobs? It'll do well on that. Like, it's heard my other talks on AI and jobs. And yet somehow I'm here. That's weird. Maybe there's something messy about it. I meet people in the green room. That's interesting for me, maybe interesting for them. You know, ChatGPT, Claude can't really do that. So I feel what I do is weird enough and varied enough to be robust. But just in general, jobs will become weirder, more unique, harder to describe. With or without AI, there'll be a lot more jobs caring for the elderly. This we've known for a long time. No reason to think that prognosis will change. In general, there will be just many more very small companies and startups. So there's already a lot of more rapid revenue growth for startups in the United States. I don't know if it's happened here yet, but it's just much easier to grow a firm when you have only, say, 3 to 5 people or 3 to 10 people, and the AIs do a lot of the work. This is already showing up in US data. These startup firms may look small, they'll have high revenue over time, they'll have a lot of ambitious projects involving just about everything in the entire world. So those projects will either hire people directly or subcontract out to have things done. And I think the optimal size for a company will shrink. Companies will feel cozier. There'll be more siblings or best friends running companies. Who you know in some ways might matter more, not in every way a positive, Well, companies will be weirder, harder to describe, less impersonal, and more based on doing something unique. So you'll just have a lot of companies that will be quite famous. You know, Midjourney would be one. They make the images, AI images. When Midjourney had its most critical breakthrough, they had only 7 employees. It's amazing you could have a world-famous company, not a rock band, not the Beatles, with 7 employees. But that was Midjourney. We're going to see a lot more of that. The people that will do best in this new world, I call them the AI maniacs. That is, they just try to learn lots about AI. It doesn't have to be on the technical side, which of course is difficult. It can just be how to use AI, how to use agents, how to automate business processes, how to integrate current human-based processes with use of AI. No one actually really knows how to do that very well, We're gonna employ a lot of people over the next few decades trying to figure that out. Pretty high failure rate. I think basically it will work, but people who are good at that and set out to do it will be the new super rich. I call them AI maniacs. They'll be orchestrators of AI. They may not be the people who build the models. The people who I think will do worse, and there's a lot of them, but those are people who play by the rules. They're smart, they come from a good family, they go, like here it would be Oxbridge, they do what they're told, they're conscientious, and they expect to walk into some kind of London life where they're a partner of something or other and earn a high salary and pay a lot of tax and decide Dubai is too warm over the summer and then wonder if that was really the right judgment after all. Those people are smart, they work hard, they're still gonna do fine. None of them are gonna starve. But their edge, just from having like this strong background and good training and conscientiousness, I think that will be in relative terms worth much less. They'll just be ordinary workers unless they somehow become AI natives. So politically, I think that will be a problem. That's a very influential class of people. I'll just close by noting a paradox. In the long run, there'll be a lot more leisure time because there'll be more wealth and AIs will do a lot of the work. But counterintuitively, in the short run, you've all got to work a lot harder. You've got to figure out AI, if nothing else. So you're like, the leisure time, the leisure dividend, where is it coming? I've never worked this hard. Well, that's part of the paradox of AI. Eventually it will come. But I would say for the next 5 years, most of you will, or at least ought to feel, you've never ever been so busy in your whole life. Thank you all for coming and listening. For me, it's a real honor to be here. And, uh, take care. Bye.
