The Creativity Code: Art and Innovation In The Age of Artificial Intelligence. – #297

The Creativity Code: Art and Innovation In The Age of Artificial Intelligence. 

For More of SuperCreativity Podcast By James Taylor

What does it mean to be creative? Is creativity uniquely human or artificial intelligence be considered creative? These are just some of the topics explored by Marcus du Sautoy in his new book The Creativity Code: Art and Innovation In The Age of Artificial Intelligence. 

Marcus du Sautoy is the Charles Simonyi Professor for the Public Understanding of Science at the Oxford University, a chair he holds jointly at the Department of Continuing Education and the Mathematical Institute. He is also a Professor of Mathematics and a Fellow of New College. He was made a Fellow of the Royal Society in 2016 and Esquire Magazine chose him as one of the 100 most influential people under 40 in Britain. In 2009 he was awarded the Royal Society’s Faraday Prize, the UK’s premier award for excellence in communicating science, and in 2010 he received an OBE for services to science. 

Technology has always allowed us to extend our understanding of being human. But will artificial intelligence actually enable us to create in different ways? And could recent developments in machine learning also mean that it is no longer just human beings who can create art? Marcus du Sautoy and I discuss this and more.

Artificial Intelligence Generated Transcript
Below is a machine-generated transcript and therefore the transcript may contain errors.

James Taylor  1:28  

I've been loving reading your new book, The Creativity Code: Art and Innovation In The Age of Artificial Intelligence.  as it's just such a fascinating topic about whether creativity is something that's uniquely human or artificial intelligence can be trained or even teach yourself to be creative. Very early on in the book, you talk about you share Elizabeth Bowden's framework for the three types of creativity, I would love to just start if you could just share those frameworks, I think it'd be quite useful as we get into the conversation, to be able to can navigate themselves. - The Creativity Code: Art and Innovation In The Age of Artificial Intelligence. 

Marcus du Sautoy  2:00  

Yes, Margaret Bowden, cognitive scientist, will thira there are two parts of this one is her definition of creativity, which is quite interesting, which is that something should be new, surprising and have value, which I think, I use as my kind of starting definition. And I think creativity is one of these words, actually, which is quite hard to pin down. It's a bit like consciousness, What is consciousness. And I think actually, the two are highly related. As I explore in the book, I think creativity is our tool for examining our consciousness. But I thought that was quite a good sort of working definition that she came up with, which has these three components, you know, novelty is something that, well, we can judge objectively whether something is new or not. And I think you know, machines can easily make things that are new, if those two other elements, which I think are really fascinating, the element of surprise, which is about engaging our emotions, making us look at things in new ways, because we sort of getting jumped out of are kind of rather a machine-like way of living. And then the last one value as well. That's very tricky because that's very fluid changes over time over geographies from one person to another. So, that's her definition of creativity. But then she has another exploration of three different sorts of creativity, which I think is really interesting, because partly, this book is as much about human creativity and trying to understand what it is we do, as much as can a machine be creative, you know, ultimately, the books about AI, and creativity. But I think throughout the book, I learned a lot about my own human creativity and her sort of three different sorts of creativity. And three, I think, are really fascinating. She has what she calls exploratory creativity, which is sort of taking the rules of the game, and just pushing them to the extremes. You know, so I think somebody like a bar, for example, is working within the Baroque period and is, is really just pushing those broad rules to the absolute extreme, then you have what she calls combinational creativity, which is sort of taking ideas from one area, and bringing those to a completely different area to give a, perhaps a new way of looking at the world. So a simple example is a kind of fusion cooking, you know, taking the ingredients from the east, but putting it through a sort of French cuisine filter. And then the third one, which is somehow the most exciting, the most challenging is transformational creativity, which is where something seems to appear out of nothing, that it's really a breaking of the system, those moments when we just see a sudden transition, a phase change. So I suppose if you think about Picasso, Picasso is that moment when suddenly things look totally different and seem like that transformational creativity. - The Creativity Code: Art and Innovation In The Age of Artificial Intelligence. 

Lovelace Test 

James Taylor  4:57  

And I gotta say as I was reading the book, I almost kind of felt your highs and lows as someone who's an expert in their field early on in the book, you start with a great story obviously about AlphaGo. And about the game of Go and that famous, you know, Lisi doll match and suddenly thinking, do mathematicians even have a job, you know? And so I can, I can almost feel your pain as you're going through the book thinking, what is our future. And you, we've all probably heard of the Turing test. But you actually mentioned something else, which I hadn't heard of this phrase, which is the Lovelace test as well. So what is that? Because I think I got a sense at different points you were going, is it passing is AI surpassing it now?

Marcus du Sautoy  5:40  

Yes. So I think, you know, creativity is a word that we've clung to is something that surely a machine can never do because it's sort of an expression to be human. And, you know, I've always used this word as a kind of protective shield against Why think machines can't do mathematics because mathematics isn't about just turning a handle on the axioms and journeying out of proof. There are a lot of choices involved. There's a lot of emotions about twists and turns and surprises. And so there is a huge amount of creativity. So this word for me has been key to saying, you know, okay, it sounds like something a computer could do mathematics. But actually, mathematics is much more human than most people realize. 

So part of the book is sort of that journey of just showing people that mathematics is much closer to creative arts than people might expect. So I thought I was safe until I saw this moment. And that's sort of the spark for the book is this match against Lisa doll of a, an algorithm that not only beats a human, and we got used to machines doing things faster, better than humans, chess in the 90s, Deep Blue beat Kasparov. But this game of Go is one that has been notoriously difficult to get a computer to play. Because it's very intuitive. It's a lot about patterns building up on this board is black and white stones, trying to surround each other. 

But it was more than just the computer managed to achieve, you know, an ability to play this game very well. It did something which I believe passes these three tests that Margaret Bowden set have value, surprise, and novelty. 

And so seeing this piece of code, do something, which was, I believe creative was what set me off on the book. But a lot of people will counter and say, Hey, hold on. But that's the humans who wrote the code, isn't it? So it's really the humans who are the creative ones. And this Lovelace test, named after Ada Lovelace who we credit as the first computer coder. She saw the machine that Charles Babbage had made to do, sort of boring calculations. And she started to creatively imagine what the machine might be able to do more. And she speculates in some notes, she wrote about how the machine might be able to play music. But she has a word of caution, she says, but we can't say that the machine is really being creative, because it's the human who told the machine what to do. So this thing the Lovelace test is kind of a bit like the Turing test, you know, can a computer create a piece of art, in whatever form music, visual art mathematics. But the key thing here is that the human who originally wrote the program cannot explain how the piece of code produced the art. - The Creativity Code: Art and Innovation In The Age of Artificial Intelligence. 

Marcus du Sautoy  8:34  

Now, with

Marcus du Sautoy  8:36  

code in the past, which was written in a very top-down manner, we were writing the instructions, which are just implemented by the machine. So I think the creativity of Ada Lovelace is right and really belongs still with the human who's writing those instructions, the machine is just implementing the thoughts of the human. 

But we've seen a real phase change in the way code is written, it's now being written in a very bottom-up manner, this thing people might have heard of called machine learning or deep learning. And this is where the code is allowed to change and mutate to rewrite itself, because of its interaction with the world around it, the digital world around it. So this means the codes are starting to separate themselves from the original human who coded it a bit like giving birth to a child, you know, first of all, the child is a combination of the DNA of the parents, but its environment quickly takes it off on an on its own track and creativity of the child. We wouldn't say oh, that's the creativity of the human. And so that's what I think is so fascinating. This line, this move that was made now famous, this creative move in this game moved 37 games to where the code suggested a move that all the humans thought was terrible at the time. But ultimately, the code managed to use this very early move to win the game. So there was the surprise that what a terrible movie, and then the value from oh my gosh, it's one AlphaGo the games Though, so, but the exciting thing is that move was considered such a bad move by humans that if any human had seen that line of code, they would have deleted it. It's really a line of code that appeared out of the learning process and therefore belongs to the code and not to the humans who originally coded it. So for me, this is a moment which kind of passes that Lovelace test, you know, something creative, okay? Not a work of art yet, but something that passes those Margaret Bowden tests, but which you can't really credit to the humans who started the code off. - The Creativity Code: Art and Innovation In The Age of Artificial Intelligence. 

James Taylor  10:34  

Now I can imagine all people hearing this, and I almost said that with you there was this. It's like having someone, maybe a competitor that's better at you than something that you can do a little bit, but I don't know, surely, surely I can. I can outpace them. But you I guess in two games further, after that, things get into a game for movie 70. At least he thought he played a movie that was so kind of creative, so innovative. And he was called God's touch or the God or the hand of God or something, they called it in the end. And it pushed the game of Go a human helping also pushed the game of Go. And in the book, you show a graph, which I diagram, which I haven't seen before. And it was a little bit of a lot of parts. In the book, I kind of got those that kind of like lightbulb moments and, and for me, it was the most optimistic part of the book. It was maybe some people have heard me that Hans Moravec, the landscape of human skills, that we're seeing AI slowly flooding this landscape and all the maybe more simple skills like arithmetic, they all went 50 years ago, machines were doing those and more recently chess, you know, logic, and then we're humans, we're trying to get to this higher ground of creativity, collaboration, those things. 

But in the diagram, and I'd like to explain that, as you call it, it's the local maximum graph. You said, actually, what these machines actually chose to do, where it's not necessary, we're not climbing just Everest, now, it's going to help us to see that there are even bigger mountains to climb, and it's going to open us up to even a new possibility, we can't even contemplate it. For me, that was the most optimistic part of the book,

Collaborator Narrative

Marcus du Sautoy  12:14  

I'm so glad you pick that up. Because ultimately, I wanted this book to be a positive take on AI, despite all of my own personal fears and society's fears. And I think, you know, too much this sort of story is the Hollywood Story of a dystopian view of AI that's going to wipe us out of talks of singularity, the moment that AI becomes more intelligent than us and then leaves us in its wake. And my feeling is that we should change the narrative that this isn't about a competitor narrative, but a collaborator narrative. And actually, I try, and I actually prefer to translate AI, not as artificial intelligence, which was always Turing's idea was, by trying to create our own intelligence in a machine will understand our intelligence that much better. So he was trying to replicate. And that's the Turing test is, can you make an intelligence artificial, that you can't tell the difference between a human but I think much more interesting is trying to create an augmented intelligence, something additional to us additional intelligence and alternative intelligence? And so I think the, you know, we have too, too much of a one-dimensional view of this where, you know, when is AI more intelligent than us? That's not the right way to look at it. I think the better way to look at it is to think well, they can do things better than some things better than us. And we can do some things better than they can. And it's the collaboration together. That means that we can go much further. And I think, what's that diagram? Absolutely, I'm glad you picked it up. I mean, what it is, is basically the idea that if you climb a hill and there's a fog around the hill, you might think that's the highest point in your landscape. But what the AI is helping us to do in a way is to clear the fog, and to show us that there's an even higher mountain and that what we thought was this, the peak of our performance, what I call this local maximum is not a global maximum. There's another Hill higher. So yeah, but you have to take the risk of going down the valley, the adaptive Valley, and up the other side. And this for me was an image that I kept on seeing throughout the book. And weirdly, you know, because we've reached that local maximum in different performative elements. We don't try anything new. And actually, weirdly, I think humans, once they find something that works, especially creatives actually, is that they just repeat the behavior because it's been successful in the past. Once you find one thing, it's a big risk to throw it away and try something completely different. So weirdly, I think that we end up quite often as humans behaving like machines, just repeating behavior. And so many stories in the book are examples of the AI showing us as humans, new ways to do things, stopping us from behaving like machines, and actually opening up our human creativity again. So that's why I see this as a really wonderful catalyst for human creativity, as much as AI creativity. - The Creativity Code: Art and Innovation In The Age of Artificial Intelligence. 

Future Ready

James Taylor  15:20  

Now, I know when you do your radio programs, and you speak a common question you must get at the end when you mean when we used to do events when it was in person, but maybe virtual now as well, is people will come up to you that maybe have children, or you've got two twin daughters, as well, and they can't do so what should my kids be studying? What skill should we be developing to be kind of future-ready? So what? What advice for any parents listening just now with these very changing places? Should they all be studying very much the STEM subjects? Or art? Or what? Or maybe something else? What should you be thinking?

Marcus du Sautoy  15:58  

Well, I think the key here is to learn about learning, not so much to learn a subject because everything is so fast-changing, you know, people talk about this as a Third Industrial Revolution. And I think they're good comparisons. But you know, the industrial revolution in the 19th century, really happened over a generation. So it was the, as you say, it was the children of the families that had to readapt. And the jobs that their parents did, were not the jobs that we're there for them as children, I think this thing is so fast-changing that a job you might have today, in 10 years will not be there. So I think the speed of this revolution is so fast that it's not only the children that need to get this message, I think it's the adults that we need to be prepared to retrain and to have the flexibility to not get stuck in just one particular discipline. So So I think the emphasis should be on developing in not just the things that you know or can do, but the ability to be fleet of foot and to retrain to relearn, and that's quite a difficult skill, but after Well, I think that was always what education should have been about. It's not so much the subject material, it's about ways of thinking. And certainly, you know, I think mathematics is a fantastic training ground for giving you that sort of flexibility of mind to be able to take a new problem and to apply tools to kind of overcome this new challenge rather than just and that's why I think, you know, mathematics often suffers from being taught in a rote way. And that's not the point. It's about ways of thinking not about being able to follow rules. So I would say, Yeah, absolutely. I think the arts are actually a fantastic tool for developing this ability to be fleet of foot and change so and you talk to so many scientists, they will talk about the impact of creative arts, in their upbringing and the importance of being a musician or a painter or writing poetry. So many scientists have that part of their lives. And for me, I think it's one of the tragedies of the UK education system or fear. It's quite a lot of places that the creative arts are really being marginalized as kind of a fluffy subject which isn't needed. And that's why I like the idea that the STEM stem is fantastic and so important. But I like this idea of STEAM where the a is all about the arts and the importance of the arts actually to making a scientist feel creative and take leaps into the unknown and take risks, that without that side, I think they become rather impotent.

- The Creativity Code: Art and Innovation In The Age of Artificial Intelligence. 

James Taylor  18:45  

I'm James Taylor, business, creativity, and innovation keynote speaker and this is the super creativity podcast. If you enjoy listening to conversations with creative thinkers, innovators, entrepreneurs, artists, authors, educators, and performers, then you've come to the right place. Each week we discussed their ideas, life works, successes, failures, creative process, and much more. 

You'll find show notes for today's episode, as well as free creativity training at Jamestaylor.me. If you enjoyed listening to my conversation with Marcus du Sautoy, then please check out my interview with Rutger Bregman Dutch wunderkind of new ideas as we discussed the case for universal basic income in the age of artificial intelligence. Here my conversation with Rutger Bregman at the Jamestaylor.me. After the break, we returned to my interview with Marcus du Sautoy, learn about the democratization of creative collaboration. This week's episode is sponsored by SpeakersU the online community for international speakers, SpeakersU helps you launch, grow and monetize your speaking business faster than you thought possible. If you want to share your message as a highly paid speaker, then speakers will teach you how just go to speakersU.com to access their free speaker business training. 

Democratization Of Creative Collaboration.

I remember visiting a Nobel Prize Museum in Stockholm and had the list of Polling exhibits there. And they have a French brewery because Barry always kept a French berry on his table. Because even though he wanted two Nobel prizes, and completed them in very different areas, he always wanted to remind himself that he was also an artist. He was a man of science, but he was also an artist as a combination. Originally, you're talking about the augmenting side. And as I was reading your book, and, and you mentioned that you have with your twin daughters as well. And I've got a friend who has twin daughters. And she said, when they started growing up, they kind of started speaking their own language with each other. And that which I'm told is quite common with twins. And in the world of technology. In aerospace, for example, we're now seeing this thing that the digital twin, you have the technology, essentially the algorithm on the ground of an engine, and they can make the changes there. And then when they find it works correctly, and they get a little bit more productivity than they'll send out to the main and the real engine. And it will do that. I'm wondering with the augmentation, do you see maybe a future that each of us as individuals in the same way that we have very simple versions, like Alexa, maybe at home, that we will have a digital twin, someone that will know as something that will know us so intimately, they can maybe take away some of the decisions we don't necessarily have to make? - The Creativity Code: Art and Innovation In The Age of Artificial Intelligence. 

Digital Twin

Marcus du Sautoy  21:18  

That's really fascinating, because, in a way, it's the theme of Ishiguro, his new book, Clara and the sun. I mean, I'm going to give it way too much for people who haven't read the book. But, you know, I think, here again, we see the importance of the arts and exploring, you know, what you've raised is an incredibly challenging philosophical idea. First of all, you know, what sort of replicate? Can you get it yourself? What, what are the moral implications of that about sort of experimenting with different versions of your life, to see which one is best? I mean, I think that it's, I mean, really, sort of fascinating challenges that that raises. But it also raises the fact I think that often what's important here is that just a very small change in your sort of surrounding conditions can concern what looks like the same person off in a completely different direction. And it's very interesting with twins because we see, then again, you know, with my twins, I've got this whole nature-nurture experiment going on. And you can see that although they have the same genetic sort of kit to start with their different life experiences are taking them in very different directions. And one of the stories I talked about in the book is an art project that was done at the serpentine Gallery in London, where an artist wanted to explore this kind of idea. And he created six, kind of virtual, what he call Bobs, and then the interaction in each gallery, they placed in six different galleries, and the audience coming in and viewing and interacting with these digital identities, which transform code, the code would update itself in this machine learning way, such that although the whole thing was still deterministic, in a way, you know, that you could follow through what the implications were, it didn't sort of some random number generator, it was deterministic, but the six Bobs that ended up at the end of that exhibition, what's so different? Yeah, they all started with the same code. 

And I think this is actually, I think, rather an optimistic discovery, because it means that, although our lives are being totally kicked around by algorithms, and I think one of the right theory is that these algorithms are just going to all funnel humanity into one direction, you know, recommend us all the same books, to read the same films to watch such that we really narrow the, what's available to us, what I'm seeing is actually the opposite that a small change in my likes compared to say, your likes on a recommender algorithm will send me into completely different bits of the sort of digital library. And so I think, this sort of chaotic nature in a technical mathematical sense, something that is very sensitive to small changes going off in very different directions, is one of the things which is making sort of algorithms actually quite interesting as kind of almost individuals and very unpredictable. Because even though this is deterministic, our point is a small change in the data that it learns from can result in it becoming very different and, and that's sort of quite exciting and frightening. - The Creativity Code: Art and Innovation In The Age of Artificial Intelligence. 

Creating Better Questions

James Taylor  24:35  

It's almost like computer DNA where you know, your DNA is part of your makeup in terms of what you end up becoming, but it's not it's not everything, necessarily not sufficient. You know, me You mentioned that art there as well. But you know, Picasso was famously quite critical of computers. He's, I think he said, he said, he said, computers are stupid or something along those lines, because they only give us answers. So do you think machines can actually help us create better questions and generate better questions that maybe will take us to new places, whether it's in mathematics or the arts or other fields?

Marcus du Sautoy  25:08  

I think that's a very interesting point. Because I think asking a question is often the most creative part of the whole process. I mean, that's often why I work with other creative artists benefits my own mathematical work because even though they can't produce an answer, they're not going to prove a theorem, they will quite often ask a question that is sort of so left field for me, and I like, Oh, that's really fun. I never thought of that before. So I think the creativity of question raising is really interesting. And you can get an AI to sort of learning from a particular database, and then make suggestions of things which can be very stimulating for the human. So one of the real challenges for AI in creativity I seen is that, although it's very good at visual arts, one of the big breakthroughs of machine learning has been about vision and computers, music as well. I mean, Lovelace I think will be quite impressed to see what music is being produced. It's the written word, where, weirdly, it's having a lot of difficulties. And I think that's because the written word is actually not just simply words in a dictionary or words in a book, it's so informed by culture through history for everything like this. But I have seen an interesting project called whim the whatever machine, which is sort of it's it's an algorithm, which is trying to raise interesting narrative suggestions for an author, what if you combine this and this, and it's sort of playing a bit on that combinational creativity that I talked about at the beginning, that I'm taking ideas and blending them mixing them? 

So it sort of learns what sort of things we like, you know, horses with wings, for example, that took over our cell phone an interesting journey with Pegasus. And so it's the whim it takes whatever genre you're kind of interested in and sort of makes suggestions. So it's a bit like question raising and what if this thing and it's clever enough that it doesn't just randomly meld things together? You know, a bit like, I used to love those picture books as a kid where you could put a random head and a middle and legs on, it would really think about you what this person is going to be interested in considering as a combination? Given that I've seen what they're interested in already, but which they haven't considered already? So? So I think that's that there is the ability of some algorithms to play that role of asking questions, even if not finishing the stories. So we'll set the person off in a new direction. - The Creativity Code: Art and Innovation In The Age of Artificial Intelligence. 

Impact Of Pandemic On Creativity

James Taylor  27:43  

Somebody also mentioned the book has lots of things I never heard of before, which is always great when you read a book or underline. One of them was the wind curve, I think it was very good. And it got me thinking. In my own work, I talk a lot about this idea of the third place, about the genius loci that places themselves can have their own ability to inspire creativity and obviously the machines. Isn't that a little bit more difficult, I guess. And I do wonder, I mean, so many the recent I speak to you and they say oh, but no, we'll the lockdowns we've been having during the pandemic, but look at all the innovations that have happened, then I can add my retort to that is sometimes Well, those vaccines are often the initial conversation that has come about from maybe a series of scientists symposium, going having a drink together in the bar or the pub afterward, and it can start we're just seeing the productivity of it just now. So I wondered, what do you think is the impact of this pandemic, especially the lockdowns of people, not physically being able to be in the same space to collaborate is having on creativity?

Marcus du Sautoy  28:50  

Yes, I mean, I think there's an interesting sort of component that the lockdowns had, which is democratizing things in a way that has given people access to that common room or to the conversations that were being held very much of you had to be in Oxford or something to be able to attend that conference and be together. So I think there's something quite exciting about giving access to people that normally wouldn't have been able to drop in on those conversations. 

But the downside is that it's all quite performative. We're, we're in this space where you know, it's that downtime. How do we simulate the downtime? That is often the most important moment of, Oh, can I have work? That was really interesting what you said in the meeting, and you want to have it offline. And somehow, when we press leave on our zoom call, often I wish that I could just go and tap somebody on the shoulder that was in that zoom call and say, really fascinating what you said there, but I didn't want to take a risk in front of everybody else. So I think we're Where, although I think there's a really positive side that many more people have been brought into the conversation. I mean, I'm, I'm part of many organizations, for example, a hay Festival, which, you know, traditionally you would have to go to this tiny little village on the Welsh English border, in order to be able to hear older writers. And we're very excited by the fact that now we can put the things online, and so many more people can come. But the downside is that the conversations that are had, even in the hall itself with a writer are always so exciting. And we just, we haven't quite understood how to create that. Yeah, the common room. Yeah, coffee.

- The Creativity Code: Art and Innovation In The Age of Artificial Intelligence. 

James Taylor  30:45  

I mean, it's, it's, I think it's also the function of video, I guess, as well, we're on video just now. But if people probably listen to this as audio, I think things like clubhouse have been very interesting, because people are having a lot of the ones that go in, especially like AI machine learning groups, people have been quite deep conversations. And the ability to be able to have a conversation with maybe someone at DeepMind or Google or who's a real specialist, and maybe you're not necessarily a specialist, is that it does feel like democratization of that, which I agree is pretty amazing. While you're talking to my hay Festival, which is a great Literary Festival, one thing I did want to how you thought about this was in the book, you're obviously mentioning different tools, different technologies, it must be as a writer, how you get that balance between mentioning certain technical tools are very net at that point now, and how you create something that's going to live a little bit longer. So I mean, I'm thinking just now, there wasn't obviously any mention of something GPT three in the book, because it just got published, probably knowing the publishing schedule, like probably written in 2018, published in 2019, or 2020. So how did you come to peace with that?

Marcus du Sautoy  31:56  

I think it was a real issue with this book, which, for example, my first book about prime numbers, you know, that's a topic we've been thinking about for 2000 years. So it's really going to survive a couple of decades. And I think that I think I was quite lucky, actually, because I even thought GPT three has come out, which is an incredible text generation. I'm already talking about text generator algorithms. And I, I haven't seen anything that really has made me think oh, my gosh, that's so different from anything that I've written in the book, I, I think I chose a sweet spot for writing this book where many of the kinds of ideas had started to bubble. And I've given accounts of sometimes early versions, but I haven't seen anything that really has been that isn't in the book. And that for me is kind of intriguing, because why is that there was such an explosion of stuff. And I think that weirdly, I think there's been a slight plateau that's happened that, yeah, we're taking different datasets and putting them in. And so I did an event with a Munich jewelry festival, looking at jewelry being morphed by AI. But it wasn't so different from things that I'd seen in the kind of visual arts realm. So I think I was quite lucky in just choosing a sweet spot where a lot of the kind of things have been tried out. And waiting, I just think we're seeing slightly more of the same sometimes better. GPT three is certainly a very good writer, but it still doesn't get over what I challenge writing algorithms in the book, which is, although it's producing the locally, very cogent paragraphs, globally, this thing just is meandering and boring. I mean, there's a book written with Kendrick, just recently a human artist and GPT. Three, and they sort of having a dialogue together. Absolutely fascinating, sort of stream of consciousness on both artists and AI is part but by the end of it, I was still although I had some stimulating thoughts throughout each paragraph in the end, and I didn't feel like I had an overarching narrative. So you had that you

James Taylor  34:06  

had the Miles Davis question of so what

Marcus du Sautoy  34:08  

he said at the very end, so you picked up on jazz because, you know, I talk about jazz improviser and I think that suffered the same thing locally, it was producing very convincing improvisation. But after five minutes, the music was becoming boring. - The Creativity Code: Art and Innovation In The Age of Artificial Intelligence. 

About Our Future

James Taylor  34:21  

So at the end of this, you went through this, you're doing all your research, seeing looking right in the boundaries of what was happening was what was your conclusion? Are you in terms of the role of human mathematicians, scientists, storytellers, and creatives? Are you optimistic about our future? Or do you think we're just playing for time now? We got another 20 years left before we're going to be got rid of

Marcus du Sautoy  34:45  

that. No, I'm still optimistic. And I think it's exciting because I think the idea is that this will be a fantastic collaborator. It is about a different way of thinking. It's making our thinking sort of multi-dimensional in a way that Without it, we're sort of thinking in just particular ways. So and, and ultimately, one of the messages is that as this thing becomes more and more complicated AI, we, as Lovelace said, you know, the Lovelace challenge, the challenge that we don't understand how it's thinking and more, we're seeing algorithms. We don't know how they're making their decisions. And that's important for society because more and more decisions are being taken out of our hands. So I think we need tools to examine the code internally. And, and as I said, in the beginning, I think creativity was our greatest tool that the human species developed, to explore our own consciousness, our own in a world, see whether my pain is anything like your pain. Some of the best examples I saw of AI art are being used to see how the code is seeing the world. - The Creativity Code: Art and Innovation In The Age of Artificial Intelligence. 

Marcus du Sautoy  35:49  

I love that

James Taylor  35:50  

I love that dissection. And I felt I almost felt like I wanted that to be another book.

Marcus du Sautoy  35:55  

Oh, yes.

Marcus du Sautoy  35:57  

As we go forward, you know, the question about whether, you know, there will be another phase change that happens, and probably far away down the line, but when AI becomes conscious, and how we can ever tell that, I mean, that's one of the unsolved problems of science, being able to tell, you know, maybe I'm in this zoom world, I've just been able to create a fantastic Avatar and actually, the real Marcus du Sautoy is, is sunning it in the garden, and I've just got this incredible avatar that you're actually talking to and convinced you're talking to a conscious human being so. So I think when it comes to when machines really pass that threshold, it's the art that the machine will produce will be our indicator, I think. So ultimately, I think that, this whole project, which at the moment is about helping human artists, perhaps to expand their repertoire. Ultimately, I think it'll be about understanding that moment when we see a real phase change in AI becoming conscious, - The Creativity Code: Art and Innovation In The Age of Artificial Intelligence. 

James Taylor  36:56  

well, Marcus, the creativity code, Art and innovation, the age of AI, fantastic book, highly recommend it. And also, even though I don't see them as strong mathematics, I really enjoyed the sections of which specifically about mathematics. I felt like I, learned as much about mathematics as I did about creativity in the book as well. So thank you for being an educator on your topic as well. Where's the best place for people to go to learn more? Not just like the book, but your other kind of work in your research? Yes, I

Marcus du Sautoy  37:24  

have a website. I'm the Simoni professor for the public understanding of science in Oxford. So one of my roles is a kind of ambassador for science. And so they go to www.Simonyi.ac.ox.uk There's a whole range of activities, things I do with the Royal Opera House with various theatre companies music and some of my mathematics as well. So that'd be a good place to start.

James Taylor  37:56  

Fantastic. Marcus, thank you so much. And I'm really looking forward to whatever your next book is going to be as well. Yes,

Marcus du Sautoy  38:01  

that was the product of lockdown is great for writers

Marcus du Sautoy  38:04  

is a good time for writers, though. So

Marcus du Sautoy  38:07  

Marcus, thank

James Taylor  38:08  

you so much for being a guest today on the SuperCreativity podcast.

Marcus du Sautoy  38:11  

Absolute pleasure.

James Taylor  38:13  

You can subscribe to the SuperCreativity podcast on Spotify, Apple Podcasts, or wherever you get your podcasts while you're there. leave us a review. I would really appreciate it. I'm James Taylor, and you've been listening to the super creativity podcast.