Please join me in welcoming Professor Daniel Suskin. Well, terrific. Thank you so much for that warm uh introduction. It is a great pleasure to be back with you all uh to deliver the first lecture in my new uh lecture series, AI Copyright and the Fight for Free Culture. And I want to begin this evening with this lecture, The Future of Creativity. And I want to dive straight in. So the concern about technology and creativity might feel uh like quite a new uh preoccupation, but it's not. Uh, it's actually as old as computers themselves. So on the screen here is a prototype for part of the analytical engine, uh, a steam-powered computer designed by the English polymath Charles Babbage in the 1830s. It's thought to be one of the designs for the first programmable computer. And Babbage is often called the father of modern computing. Now, this is Ada Lovelace, one of the great British mathematicians, often said to be the mother of modern computing. And she studied Babbage's machine and in 1843 wrote this. The analytical engine has no pretensions whatever to originate anything. It can do whatever we know how to order it to perform. Now, this idea of uh Ada Lovelace is well known today because of this man, Alan Turing, um of the great British computer scientists, who in 1950 uh published uh an article in the journal Mind, one of his most famous articles, entitled Computing Machinery and Intelligence. And in this article, he writes about what he calls Lady Lovelace's objection. And he says her objection was that machines can never do anything really new, that machines can never take us by surprise. In short, Ada Lovelace's objection is that AI, or machines, can never be creative. So I want to begin with this observation that this thought that machines of any stripe could never be creative, it was there from the very beginning. It is not a new idea. But just as there were those in the very beginning who said that machines could never be creative, there were also those in the very beginning who said that they can be creative. So in 1955, John McCarthy, Marvin Minsky, Nathaniel Rochester, and Claude Shannon sent a proposal to the Rockefeller Foundation for the Dartmouth Summer Research Project on AI. And this is what they wrote in that now famous proposal. We propose that a two-month, ten-man study of artificial intelligence be carried out during the summer of 1956 at Dartmouth College in Hanover, New Hampshire. The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can, in principle, be so precisely described that a machine can be made to simulate it. We think that a significant advance can be made in one or more of these problems if a carefully selected group of scientists work on it together for a summer. Now, I love the hope and optimism of this proposal, and they received the funding for what has become known as the Dartmouth Conference. Here they are gathered together in 1956 on the lawn outside the Dartmouth Mathematics Department. Many people now believe this is when the field of AI really began. It's when the field is given its name, artificial intelligence. And one of the things, one of the aspects of learning, of human intelligence that they thought would be possible to be done by AI was creativity. One of the things they set out to do that summer, if you look at their original proposal, was to build an AI that could act creatively. Now, for those of you who know the history, uh as it happened, no particular advance worth celebrating was made that summer. But nevertheless, a community formed, a direction of travel was established, and a handful of some of the greatest minds of that time began to work together. And in time, what we would see is that an eclectic collection of very different problems would be swept together under the banner of artificial intelligence. So again, from the beginning, there was this tussle in AI. Those who were saying that machines can never be creative, and those saying that they can. And this is what I want to explore with you today. Um, where we are in this debate, this long debate about technology and creativity, and what it means for the future. And in particular, this evening I want to do six different things. The first is I want to look at some of the early moments, the first signs of AI encroaching on tasks that require creativity from us. Then I want to look at a turning point, the first moment where I think it really becomes clear that these tasks, those that require creativity from us, are now within reach. The third thing I want to do is a realization, is explore a realization that has unfolded in the last few years. That asking the question, can a machine be creative? It seems to me to be the wrong question to be asking when these machines behave or perform very differently from us. Fourthly, I want to come to the present and in particular how to think about the generative systems that have arrived in the last few years. The point I'll make here is that it is a remarkable chapter, but it is a remarkable chapter in a far longer story. And then fifthly, I want to look to the future, think about what this might mean for the future of creativity, bring all these ideas together. And finally, in closing, I want to set out some implications, bring these ideas back down to earth. Uh, and it's these implications that are going to be the focus of the rest of the series. So let's begin with those early moments. And the first signs of AI encroaching on tasks that require creativity from us actually happened uh you know many decades ago. The truth is, if you look carefully, there are many pretty early signs that it's underway, and often in parts of our lives that you would perhaps least expect to see these systems doing creative things. So take writing, for example, creative writing. Here is Christopher Stratchy, one of the great early British computer scientists, sharing the love letters that his system had written in the early 1950s. Uh, MUC here, uh, the signature for this particular love letter. Um, Manchester University Computer, one of a series of computers, some of the most powerful in the world at the time. Darling sweetheart, you are my avid fellow feeling. My affection curiously clings to your passionate wish. My liking yearns for your heart. You are my wistful sympathy, my tender liking. Um there's been some improvements uh since, but you can catch a glimpse, I think, in there of something quite interesting. Or take the world of music. In 1956, Lajaran Hiller, uh, a chemistry teacher at the University of Illinois Urbana Champaign near Chicago, uses the school's only computer, Iliac uh 1, short for Illinois Automatic Computer, the first supercomputer in an academic institution, to compose this piece of music, um, a string quartet. And if anybody wants to have a listen, do share the image with one of the generative systems, and I'm sure it will be able uh to play it for you. Or take the world of art. In the late 1970s, Harold Cohen used a program called Aaron to generate a drawing. And then he transferred it to Lithograph and added colour by hand. So in the second half of the 20th century, that's just a tiny glimpse. There's a sprinkling of quite interesting initiatives, each of them in different ways, testing and probing the domain of creativity, starting to ask of us is this necessarily a uniquely human activity? But in the 1990s, these disparate efforts start to coalesce into something that looks much more like a formal field. And it's the field of computational creativity. Now, the late Margaret Bowden writes this in 1990. It's a completely fascinating book on creativity and machines. I think it's one of the best books, even still, but certainly one of the best books of that moment. So anyone with an interest, do take a look. Um, the field gradually taking shape as the 1990s begin. One of the most interesting moments, though, happens in the late 1990s. And the context for uh this moment is this, which is that in 1979, one of my uh intellectual heroes, a man called Douglas Hofstadter, writes this book, Girdle Escherbach, and it's a cult book at the time. And in this book, he's a sort of, he's a kind of fascinating sort of polymath. Um, and he is very skeptical of the idea that AI could ever be creative. And one of the things he writes in this book is this he says, a program, he's thinking about music and the idea of an of AI ever being able to generate music, and he says, a program which could produce music as they did, in other words, um as a human being did, would have to wander around the world on its own, figuring its way through the maze of life and feeling every moment of it. It would have to understand the joy and loneliness of a chilly night wind, the longing for a cherished hand, the inaccessibility of a distant town, the heartbreak and regeneration after a human death. It would have to have known resignation and world weariness, grief and despair, determination and victory, piety and awe. Piety and awe. Now, the point here is that he is, as reflected in that quotation, skeptical of the idea that these technologies will ever be creative in the way that human beings are. Nevertheless, in the sort of, you know, reflecting the kind of intellectual curiosity that he was famous for, in 1997, he gathers together at the University of Oregon a really interesting group of uh students and people to listen to three pieces of music. One of the pieces of music is by Bach. One of them is written by one of the academics at the institution, a man called Dr. Steve Larsen, who taught music theory at the university, and it's written in the style of Bach. And one of them is composed by a computer program called EMI, which stands for Experiments in Musical Intelligence. And the man here is David Cope. Um, and in the 1980s, he writes about how he suffered from a particular type of writer's block, what he called composer's block, while attempting to put together an opera. And so he turns to his computer, and the program that he builds, Experiments in Musical Intelligence, EMI, was the result. So the audience, this group of students who are listening to this piece by Bach, the real piece by Bach, this piece by Steve Larsen in the style of Bach, and this piece generated by EMI, a machine in the style of Bach, the audience don't know which is written by which. What's the result of this experiment in 1997? The audience thought that the piece written by Dr. Larsen was the one written by the computer. And the piece that was composed by EMI, or ME as it's pronounced, they thought that was the one that was in fact written by Bach. So already, you know, three decades ago, there are sort of interesting tremors unfolding at the frontier of AI and creativity. But in general, I think the most important thing to note here is that people in the spirit of Hofstadter were relatively skeptical about AI and creativity, in spite of all the progress that was unfolding. And you can see this skepticism more generally. So a good example of this skepticism about creativity ever being within reach of machines is Gary Kasparov, who in 1988 makes the claim a computer will never be able to beat Karpov or me. Now, of course, we know that that was a mistake. In 1997, Gary Kasparov, of course, sits down with Deep Blue, a computer system owned by IBM, and is beaten. What's interesting about this though is that two decades later, when Kasparov writes his autobiography, reflecting on that match and what he has learnt about AI and creativity since then, his view is that the limit of machine capabilities, of AI, is human creativity. The subtitle for that autobiography, Deep Thinking, is Where Machine Intelligence Ends and Human Creativity Begins. A sense that human creativity marks some impassable uh frontier beyond which AI cannot encroach. And again, this is a a general view during you know uh during this the start really of the 21st century. But there is a really interesting turning point, and it's a turning point that I've referred to in the previous lecture series, but I want to explore it again in this different context. And it's a turning point that comes in 2016. And it comes, of course, at the Go board. Now, I don't play Go myself, but what I know about it is that it's combinatorially incredibly complex. So in the game of chess, the number of possible moves you can make at the beginning is relatively well bounded. There's only a certain number of moves you can make. Whereas in Go, there are a huge number of moves that you can make in the first move, and then on the second move, a huge number of moves. And so the game very quickly could take many possible paths as a sort of explosion in possible paths. And that is why, back when Gary Kasparov lost to Deep Blue, figures like Pete Hoot, a computer scientist at the time, said it may be a hundred years before computers beat humans at Go, maybe even longer. The general sense was that the game of Go, again, was just too combinatorially complex. Chess in comparison was a simple game, but there was no prospect of an AI beating a human Go champion at the board. But of course, in 2016, that is exactly what happened. The then world Go champion, a man called Lee Sadol, sat down with AlphaGo, which was a system owned by DeepMind, which in turn is owned by Google, and it beat him four games to one. Again, remarkable moment and interesting in its own right. But what was particularly interesting about this game, and again, it's a move, it's a moment I shared before, but I want to revisit it, was a particular move in the second game, the 37th move. Now, I don't play Go, but I am told that there is a rule of thumb that any beginner at the game knows, which says, do not put a piece on the fifth line from the edge. So you can see the board is divided up into lots of vertical lines, lots of horizontal lines. There's a rule of, and you put the pieces on the intersections of these lines, there's a rule of thumb that supposedly any beginner knows that says, do not put a piece on the fifth line from the edge. And yet, in the 37th move in that second game, that is exactly what AlphaGo did. And most importantly, it went on to win the game. And I was watching the game live at the time, and it was remarkable because the commentators were speechless. They couldn't explain what had happened. One former champion called the move beautiful. Another said it brought tears to his eyes. Lee Sidol uh himself put his head in his hands and actually had to step out the room shortly afterwards onto the rooftops to look rather wistfully over Seoul. Um this is a still from a terrific documentary on the AlphaGo victory, which is available on YouTube. Again, I encourage you to have a look. It's a wonderful, uh, a wonderful uh uh story, very well told. Why am I, and I should say this move was so um so significant that uh Google have in fact named their new office, you might have noticed in King's Cross. If you come out the station at King's Cross, you see that huge office. It's named Platform 37 after this 37th move in the second game. Now, why am I going on about this move? And why have Google named their building after this move? Well, it's because if we had seen a human being play that move, we would have said, gosh, isn't that creative? Um but, and this is the really important point, it just feels wrong to say that what AlphaGo did in that move was creative. Creative feels like too human a word. Was it an original move? Yes. Was it a surprising move? Yes. The commentators were speechless, but was it creative? I want to plant the seed of this thought that creative is the wrong word to describe what exactly this system did in the 37th move at the second game. And I think there is something deeper going on, which that move in the 37th move in the second game is a good example of it, but of something more general. And the reason I put to you that I think it feels wrong to call what that machine did creative, is because creativity feels like a very human word. It feels like a human faculty. But this machine, AlphaGo, was working in a very different way to us. It was using enormous advances in processing power and data storage capability and particularly algorithm design in the last five to ten years, um, to perform a task that might require creativity from us, but it was performing this task in a very different way to us. Um, and this observation that the system was, in a sense, performing this task, namely coming up with something original, surprising, novel, but doing it in a different way to us, is a is an observation I want to spend a little bit more time just unpacking. And again, followers of my work will be familiar with this idea, but I think it's a really important one. So if we go back to Dartmouth, to that gathering in the late 1950s, where these computer scientists and philosophers and thinkers come together in the field of uh to found the field of artificial intelligence, there was a tacit assumption among this group. What these researchers all had in common was that there was one fundamental solution to all the different problems that they were interested in. Not just the problem of creativity, but all the different problems that they were exploring under the banner of artificial intelligence. And this is the tacit assumption that I think they were being guided by. That building a machine to perform a given task meant observing how human beings perform the same task and copying them, imitating them, replicating them. It's a belief that artificial intelligence had in some sense to ride on the coattails of human intelligence. And this was true for those working in 1956 at that Dartmouth gathering, but importantly, I also think this has been true for AI researchers for almost all of the 20th century that has followed. Why have AI researchers thought that in order to perform a task means observing how human beings perform the task and then copying them? Why have AI researchers thought that way? I think there are two reasons. The first reason is that human beings until recently were by far the most capable machine in existence. So why not try to build machines in their image, albeit machines built of silicon rather than of flesh? So that's the first reason. These systems are just the most capable. Human beings are the Most capable? Why not try to copy the way that they think or reason? Or indeed, even to some extent, the anatomy of the human brain. But there's also another reason as well, which is that many of the AI researchers until recently thought of themselves as cognitive scientists, not computer scientists. They were imagining they were working in a subfield of a far bigger project, which was understanding the human brain. And for many researchers, the prospect, the project of understanding human intelligence for its own sake was simply a lot more interesting than merely building capable machines. That was the central purpose of artificial intelligence, to better understand human intelligence. Or as the late philosopher John Searle put it, the only purpose of AI was to act, quote, as a very powerful tool in the study of the mind. And so you see this in the writing at the time. If you look at their writing, it's full of excited references to classical thinkers and their reflections on the human mind, people like Leibniz and Descartes and Hume. These were the figures in whose footsteps AI researchers thought they were following. They were thrilled by questions not about machines, but about human beings. What is a mind? How does consciousness work? What does it really mean to think or understand? So just to see this in a practical sense, think of two of those Dartmouth attendees here, Herbert Simon and Alan Newell, some of the founding fathers of the field, present in 1956 at that Dartmouth conference. A few years after the Dartmouth conference, they turn to the process of creativity in a Rand paper, which was published in 1958. And if you look at this paper, and if you look at how it is they are thinking about the idea of creativity, they write, we would have a satisfactor satisfactory theory of creative thought if we could design and build some mechanisms that could think creatively. Or I've misspelt that somewhat uncreatively. Creatively. Exhibit behavior just like that of a human carrying on create uh creativity of creative activity. Exhibit behavior just like that of a human. Again, there is a sense that for a machine to be creative, it has to, in some sense, copy human creativity. That it's got to copy or imitate the way that we think and reason when we're behaving creatively, to follow the rules that we follow, or indeed, again, the anatomy of our brains. But again, think of 2007. This system wasn't trying to copy the way that Lee Sadol thought or reasoned. Again, it was using remarkable advances in processing power, in data storage capability, and in algorithm design to perform this task, coming up with clever moves at the board, in a fundamentally different way to him. Or again, think of Gary Kasparov in 1997. This system wasn't trying to think or reason like Gary Kasparov. It was able, because of the remarkable technological advances of the time, to calculate up to 330 million moves a second. Gary Kasparov at best could juggle 110 moves in his head on any one turn. He was blown out the water by brute force processing power and lots of data storage capability. Again, the system was playing this game, the game of chess, in a fundamentally different way to him. My favorite example of this idea that machines do not have to copy our faculties in order to outperform us at solving problems that we use those faculties to solve is um Watson. And in 2011, um its claim to fame, system owned by IBM, its claim to fame was that it went on the US quiz show Jeopardy and beat the two human champions at the game of Jeopardy. And I, you know, it's an amazing moment. But what I love about this, looking back on it almost a decade and a half later, is that the day after Watson won On Jeopardy, the Wall Street Journal ran a great piece by the late philosopher John Searre, who I just mentioned before, with the title Watson Doesn't Know It Won on Jeopardy, right? And it's brilliant and it's completely true. You know, Watson didn't let out a cry of excitement, it didn't call up its parents to say what a good job it had done, it didn't want to go down to the proverbial British pub for a drink. The system wasn't trying to copy the way that those human contestants thought or the way that they reasoned, but it still outperformed them. And so I want to come back to this observation I made at the start that asking, can AI ever be creative? In light of the technological developments that are now underway, I don't think it's the right question to be asking. In fact, the real question, it seems to me, the most important question in a world where AI performs tasks very differently to us, is can AI solve the problems that human beings use their creativity to solve? In a world where AI performs tasks very differently to us, can AI solve the problem that human beings use creativity to solve? So the natural question is: to what problem is our creativity the solution? Why do human beings deploy their creativity? And in my view, to come right back to the start, to those conversations between over time between Ada Lovelace and Alan Turing and others, it's one of originality. When we want something original, when we want something novel, when in the words of Alan Turing, we want to be taken by surprise, we go to our fellow human beings and we say, look, I need you to be creative. I need you to come up with something, I need you to be original. Can machines, you know, is it right to say that machines can never take us by surprise? Again, go back to that moment in 2016 when Lee Sadol was playing the game of Go. The commentators were speechless. Very few people at the time were able to explain what had happened. It was a move that took everybody by surprise. But was it creative? Again, I feel like creativity is the wrong word to describe what these machines were doing. They were performing the task that we use creativity to solve, namely one of originality, novelty, taking one another by surprise, but doing it in a fundamentally different way. So where are we today then? Ten years on from AlphaGo's victory, from that turning point. There is now, it seems to me, a relatively steady stream of people using generative AI, whether it's ChatGPT from OpenAI, Claude from Anthropic, Gemini from Google, or Grok from X, using these generative AI systems to do original things, to generate original text, original images, original video. Are these systems being creative? I would argue I don't think so. I don't think they are being creative. But they are solving a problem that we use creativity to solve, namely one of originality, but by performing that task in a very different way to us. In 2016, and to some extent in 1997, we caught a glimpse of that when AI was playing games like Go and chess, but today we're starting to see it in so many other ways. AI writing poems, composing music, designing buildings, solving hard problems in mathematics in beautiful ways. Another way to do to think about this is go back to that description of the field of computational creativity that began to emerge in the 1990s when people in computer science started to take the idea of machine creativity seriously. In light of the arguments that I've made, in fact, the the title computational creativity was probably the wrong title to give that fledging field. In fact, I think a better title for that field, and indeed for what we have today, is probably computational originality. These are systems using remarkable progress in computational power, data storage capability, and algorithm design to do original things, but doing it in a way that doesn't look a lot like the faculty of creativity. So let's think a little about the future. What do these technological changes mean for the future of creativity? I think one important observation to make here, and this is the one that has been preoccupying me quite a lot in the last few months, is in the 20th century, I think we became pretty accustomed to the idea that the most original, novel, surprising ideas came from the heads of smart human beings. I wonder, I suspect, that in the 21st century, the most original, novel, surprising ideas are going to increasingly come from these technologies instead. And I think we can already see this to some extent in the field of scientific discovery. So think of what is happening at the moment in science. Traditionally, we have thought of great ideas about the world, great scientific discoveries coming from the heads of smart human beings. Isaac Newton resting under the apple tree at Wallsthorpe Manor and dreaming up the idea of gravity, Albert Einstein reviewing applications at the Zurich Pattern Office and stumbling upon the idea of relativity, Werner Heisenberg striking out alone on the bleak island of Heligoland in the North Sea and coming across the idea of quantum mechanics. This is the kind of romantic idea of a human being as the sole generator of original, novel, surprising ideas. But that is starting to change. So another recent AI developed by DeepMind, this one called AlphaFold is, I think, an interesting case in point. So in 2020, it solved the so-called protein folding problem, a long-standing problem in biology. There are millions of proteins out there in the world, and if we want to understand how disease works and how to treat disease, we need to know the 3D shape of proteins. But it's very difficult to figure out the 3D shape of proteins. And until recently, we knew maybe 15% of the 3D protein of the 3D shape of proteins out there in the natural world. It's said to take up the entire PhD of a good student in computational biology to figure out the shape of just one of these proteins. You know, a really labor-intensive task. AlphaFold in 2020 stepped forward and essentially solved the protein folding problem, figuring out the shape of almost all of these proteins and making them available for researchers online. It was once thought of as being the Fermat's last theorem of biology on account of its fame and its immense difficulty, and now the system has essentially solved it. What's interesting though is that Demis Hasabis, the co-founder of DeepMind and one of his colleagues, John Jumper, won the Nobel Prize in chemistry for developing Alpha Fold. But there is part of me that thinks part of that Nobel Prize ought to also go to this system as well. You know, there is a sense in which it is this system that is partly responsible for solving the protein folding problem as well. And for those of you interested in mathematics or the frontier of mathematics, something similar, extraordinary is now underway. We are seeing more and more of the most difficult, intractable, long-standing problems in mathematics being solved by these AI systems. So this in May of this year, the planar unit distance problem, first proposed in 1956, possibly the best-known problem, outstanding problem in combinatorial geometry, was solved by OpenAI a few months ago. A few weeks ago, OpenAI announced they had potentially solved the Navier Stokes Millennium Prize problem. So this is one of the most important outstanding problems in mathematics. The Clay Mathematics Institute offers a $1 million reward for anyone who is able to solve this problem. This is, I think, a glimpse of something that is going to spread far more widely through our working lives. At the moment, it is unfolding particularly conspicuously in the world of science. But I think we are going to see more and more of this elsewhere too. Where the very where the most original, where the most novel, where the most surprising, where the most consequential ideas come increasingly not from the heads of creative human beings, but from these increasingly capable technologies as well. And I think, just again, this is true not simply for scientific discovery, but for creativity too. If AI can win Nobel Prizes and Millennium Prizes by coming up with original, novel, surprising solutions, they are going to be able to write original marketing copy. They are going to design surprising buildings. They are going to be able to write novel stand-up routines and so on. They are going to be able to do more and more of the things that we use our creativity to do. So what are the implications of this? What does this actually mean in practice? I think there are a set of really important implications for thinking about what these technological changes mean and for thinking about the future of creativity. One implication is just really practical for those building the machines, for those working in computer science. That just as an aeroplane doesn't need to flap its wings like a bird to fly, just like a submarine doesn't need fins and a tail to swim, AI doesn't need to think or reason or feel like us, or indeed be creative like us, in order to solve problems that we use our faculties to solve. Again, just to really emphasize this point, it doesn't need to be creative in order to do original, novel, surprising things. Part of the consequence is also psychological. I think there is a collective psychological reckoning for us now, too. And you can see it in the field of computer science if you look at the history. So this is Douglas Hofstadter reflecting on um on how he felt when he saw Emmy in action, uh, deceiving the audience of students at the University of Oregon. I was terrified by Emmy, he said. Terrified, I hated it, and was extremely threatened by it. It was threatening to destroy what I cherished about humanity. I think Emmy was the most quintessential example of the fears that I have had about artificial intelligence. I think we have to come to terms with the idea that human beings just might not be so special when it comes to creativity. I think history is punctuated with moments that have forced us to question what it really means to be a human being. You know, when Copernicus realized that the Earth orbited the Sun and not the other way around, we had to make sense of the fact that we no longer sat at the center of the universe. When Darwin discovered the idea of evolution by natural selection, showing that we were distantly related to soft-bodied, soft-bodied sea slugs that lived hundreds of millions of years ago, we had to accept the fact that we were not divinely chosen. And when Freud revealed the hidden power of our subconscious, and that we were not, in fact, the ruthlessly rational creatures we had imagined, we had to process the fact that we knew far less about our opaque minds than we thought. And today I think AI is forcing a similar reckoning upon us. I've been writing and thinking about the impact of technology, and particularly AI, on work and society for the last 10 or 15 years. And time again during that time, people have said to me, ah, but creativity. That's the sort of thing a machine will never be able to do. I think we might have to come to terms again with this idea that while machines might never be creative like us, they might be able to solve problems that we use creativity to solve, but do it in a very different way. It has implications for my own profession, economics. It means I think economists, all of us, have to revisit assumptions about the capabilities of technologies and what that means for the future of work. This was the focus of the first series of lectures that I delivered, but just let me give you a flavor of what I mean here. So this is a really canonical paper, one of the most important academic papers in the literature exploring the impact of technology on the labour market, published at the turn of the century. And in it is this assumption that tasks demanding flexibility, creativity, generalized problem solving and complex communications do not yet lend themselves to computerization. Models based upon that assumption that creativity is somehow out of reach, I think, need to be revisited. Or take this paper, again, one of the most important papers in economics in the last few years, exploring the impact of technology on the labor market, the future of employment, how susceptible are jobs to computerization. It was written in 2013, so a decade after that previous paper, published a few years later in 2017. Again, crucial assumption here. Seems unlikely that occupations requiring a high degree of creative intelligence will be automated in the next decade. Again, you know, this was the subject of my previous lecture series, and those interested in thinking about what the impact of these changes might be for the world of work, I'd redirect you there. I think there are also really important implications now for the creative industries in particular. The group of people who define themselves by the very faculty of their creativity. Struck by this article by Amul Rajan from uh about a decade ago. Be creative if you want to outsmart the robots. I think given the technological changes that are taking place, that is not entirely obvious as a piece of device any longer. Or Andrew Lloyd Weber and Alistair Weber just last year. AI can replicate patterns, but it does not create. Well, it might not be creative like us, but it can be original. It can take us by surprise, it can do novel things. Or James Cameron, the great filmmaker. What generative AI can't do is create something new that's never been seen. I would argue, on the contrary, that we see lots of examples of these systems doing precisely that. We might not call it creativity, but it's certainly able in a whole variety of domains to come up with things that are original, that are novel, that are surprising. And finally, and most importantly, in my view, these technological changes have implications for copyright. So today, intellectual property law, the collection of formal rules and regulations that protect people's rights to their intellectual creations, is the most important toolbox that societies have to shape the creation and distribution of new original ideas. It's really a deceptively dry term for one of the most profound questions that we can ask about how we live together in society. Who actually gets to own and control the ideas that we create? Now there is, in response to the realization that these technological changes are putting immense pressure on copyright, huge disagreement and an intense debate bubbling up. So here is a statement on AI training signed by over 50,000 people from creative industries, including many of the great and the good. The unlicensed use of creative works for training generative AI to generate original, normal, surprising things, is a major unjust threat to the livelihoods of the people behind those works and must not be permitted. I think the sorts of fears and concerns embodied in this sort of statement are entirely understandable. I completely understand what it is that people are worried about. And this is the debate that I want to explore in depth in the lectures to come. There's many different ways we can think about this debate about who gets to own and control ideas in society and how we might respond to it. But I think a really useful framing is perhaps the simplest. And it's perhaps also the most provocative framing I could choose. It's a framing that the great legal uh scholar Larry Lessig used in a different debate when he was reflecting on the role of copyright at the start of the internet era. Um, and that is a moment I'm going to return to in later lectures. But he framed this debate as a conflict between the old and the new. And when he framed it in this way, he quoted the great political mind Machiavelli. And he quoted this line from the prince. He said, and this is Lessig quoting Machiavelli: innovation makes enemies of all those who prospered under the old regime, and only lukewarm support is forthcoming from those who would prosper under the new. Their support is indifferent partly from fear and partly because they are generally incredulous, never really trusting new things unless they have tested them by experience. So I want to finish with that provocation for you all, this provocative thought about the tension between the new and the old, and encourage you to join me in the lectures that follow to explore, as I say, what I think is one of the great questions that we face at the moment in this setting, which is who gets to own and control original, novel, surprising ideas of society? So I will finish there. Thank you very much, everyone, and I look forward now to some questions in the future.