Der Kreativitätscode - Marcus du Sautoy

He's becoming quite a hard act to follow these days, actually. I think he's going to put

me out of a job with all his line-up of jokes. How many mathematicians do we have in the room

just to give me a sense of who I'm talking to. The mathematicians in the room, you probably

have a similar experience that I do when I go to parties and you get this question about what

you do. Perhaps you all fake it. I'm starting to fake it, actually, and say I'm the international

spy or something. Most of the time, I kind of stock set of reactions. One of them is that

they just flee to the other end of the party and I'm abandoned. Before they go, they always

tell me what they've got in the GCSE or O-level, which is not really quite sure what that's

about. But if they do stick around, one of the things they often say is he's trying to work out

the tech here, which is doing a talk about tech is always a disaster, actually using the tech. One of

the questions I often get are statements. Come on, surely you must have been put out of a job by

computer by now. I think it's mostly because people think that what I do in my office up here is

long division, lots of decimal places. If that were true, certainly my computer would have put me

out of a job. We're all, as Alan said, we're all a little bit threatened at the moment about

this advancing AI that it seems to be very powerful, doing lots of interesting things, and surely

aren't computers all about logic and mathematics. It wouldn't my job be one of the first to be

threatened. I got this a lot during the 90s, actually, when a deep blue beat Kasparov, because

often people used to compare the idea of playing a game of chess to doing mathematics. There are

certain logical news you can make with the pieces. There's a kind of end game that you're after,

the end of the proof, the QED, where you sort of win the game, or you win the proof. And so a lot

of people said to me during the 90s, well, come on, you must be next. But I never felt

particularly threatened by chess. Actually, there was always another game that we mathematicians

always use as our kind of protective shield against the idea that computers could do our subject.

Because there's another game, and that's the ancient game of Go. This Chinese game played on a

19 by 19 grid, where you put black and white stones down. You try to engulf the other person's

territory before they engulf yours. And this is a game which has a high degree of complexity,

and much more complex, actually, than chess. There's a lot of pattern recognition that needs to

go on to be able to play this game. And that's something that really mathematics is about.

It's about spotting patterns, underlying patterns. But quite often when you're playing Go, and

especially when you're doing mathematics, to be able to quite know where you're going requires

a lot of intuition, a lot of creativity, not quite sure why you're making moves. You spend a lot of

time in this world, and you build up a feel for playing Go or doing chess. And very traditionally,

in computer science lectures, they would always say, yeah, chess is something we can automate,

because we can understand the kind of logical implications of playing particular moves. We can

follow the tree of possibilities through. Go was always said to be a game that no computer would

ever be able to play. And certainly, any attempt of somebody trying to encode the way that

human plays this game always failed. The attempts to try and encode playing Go in a some sort

of algorithm wouldn't even beat an amateur at this game. So I felt pretty safe, because computer

science say you can't play Go. So you certainly won't be able to play the complex game of

mathematics. So I got a little bit of a shock a couple of years ago, and you're probably aware

of this story. When a team in London declared that they had got an algorithm that they believed

could compete at, not just a high level, but the highest level. This is DeepMind in London.

And actually, it was Demis Hassabiss who went to see, he went to Cambridge to do his study,

and he was told this old adage that you can't play a programmer computer to play Go.

And this was like a red rag to Demis. And so he went away and set up this company,

and they devised this algorithm that they thought could play the best. And they challenged Lisa

Dall, a Korean grand master. And Lisa Dall was totally dismissive of this algorithm that wouldn't

be able to get anywhere near the level that he could play at. He said, I'm going to demolish this

thing five nil. They were going to play over five games. But he got a little bit of a shock.

I sat and watched these games obsessively on YouTube because I realized that my life was

probably under threat. And as I watched, I saw Lisa Dall get more and more depressed throughout the

games. He lost the first game. He lost the second. He lost the third. He lost the match already after

three. He won the fourth game. And he now regards that as the greatest game that he's ever played,

that he was able to beat this algorithm in one game. And he lost the fifth game. So he lost

four one. What had changed in the last two years? The style of coding has changed. And we've got

new sort of code on the block, which is able to do things that code in the past couldn't do.

And it's something that we're very interested here in Oxford, this idea of deep learning or machine

learning. So code in the past used to be written in a very top-down manner. You really had to know

what the thing was, how the thing was going to behave. You told it the rules of how it was going to play.

You had to understand the setting. And the machine just implemented that. Sure, it could play

chess because it was told to implement this thing. It could go deeper. It could analyze more

situations than a human. But the human was still telling the program what to do. What has changed

is that the code is now written in a very bottom-up manner. We've got a sort of code which is

learning very much like a child learns. In the past it was like the parents DNA would give birth

to a child, but the child would be still attached to the DNA of the parent. It wouldn't learn

anything new. But suddenly we've got code that can adapt and change and mutate and re-parameterize itself

as it encounters a new environment. There's almost like a meta-code which is telling the code

how to change if it gets something wrong and change and mutate. And this is what they use to actually

train AlphaGo to learn how to play this game by playing games and failing. But they started with

some other games to start with. Some simpler games. So in fact, they started with Atari games.

Games are used to be obsessed with actually when I was a kid. And one of the ones I really love

with this one called Breakout where you have a little a ball which ping ponges up and down. You've

got a paddle and you've got to knock these out and you score points. The blue are just one point

up to red the higher points. The machine was only given the pixels on the screen and the score.

It had to learn how to play this game. It wasn't told anything about the fact that you had to hit

this ball. But it was randomly moving the paddle and every time it hit the ball it saw the score go up.

So it re-parameterize itself and said I'm going to prioritize moving towards where the ball came

now when I was with my mate and we played this we were very pleased when we found a fantastic hack

because you can create a little tunnel on the left hand side and then if you get the ball to go up

there then you don't have to do any work at all because the ball just bounces backwards and forwards.

What I was absolutely staggered when you see how this deep mind actually you learned to play

the Atari game. It learned the same hack. Not just humans but the computer is also lazy. He doesn't

want to move this thing around. So shoots it back up again. This is extraordinary after 600 games.

It had learned just by randomly moving the paddle and seeing that what the moves were that made

the score go up fast. It had learned how to do this hack. So this is how it then went on to play

the game of go. So what it did was to first take all the human games that are on the internet.

A lot of games encoded on the internet. It learned how humans lost and won the games and that was

its first material that it learned on. Then it started to create synthetic data. It started to play

itself. Different versions of itself would play a game and then if it lost a game it would understand

which were the moves that meant that it actually was failing to play at a high level and it would

re-parameterize itself. So after a while it got to this high level. Quite amazing that it did

such that it could challenge Lisa. Now first sight you say okay so the community has just got

very good. It can analyze its game very deeply. But I think that there was something

more amazing that happened during these games. The first game interestingly, Lisa Daul decided

that if it had done learning on how humans played maybe the best way to beat this computer

was to play rather unlike a human. So he actually played a very disruptive game but the

AlphaGo was smart enough to actually just cope with the moves that he was making and they

turned out to be quite weak moves and he lost the game because he was not playing his standard game.

So in the second game Lisa Daul decided to play a much more standard high level game that he knew

very well. Very early on in the game your go master teaches you a few things strategies. One is

that you should play on the kind of edge of the board. So you're encouraged to play on the first

second third and fourth rows in because there's a kind of competition early on for the edges

and the kind of internal part of the board. And if you play too far into the middle of the board

early on it's considered a weak move because you're not really establishing important territory

at that point. So Lisa Daul on the 36 move of this game decided that he needed a cigarette

break and he went up to the top of the hotel had a cigarette. AlphaGo didn't need to smoke in

order to get stimulation. So it's that there and it's thought for a while and then it asked the

human player because this wasn't an exercise in robotics this was an exercise in just pure thought. So

there was still a human this is actually still quite difficult for an AI to actually pick a stone up and

place it on the board but it told the human player to place a stone on the fifth row in. So I've

circled this he was playing he already anthropomorphized the AI. The AI is playing black and it put

this stone in on the fifth row in which I put a little white circle on. All the commentators are

a member this on YouTube. They all gassed and went whoa wow. AlphaGo has made a huge mistake

never play that sort of move early on in the game. This is a Lisa Daul will be in it will be one

one after this and they were all very complacent and Lisa Daul came down after his cigarette break.

Look at what AlphaGo and you can see you should watch this bag it's so funny because it just

Lisa Daul like cannot believe what this AI is just suggesting what a stupid move.

But he's a bit more suspicious of things you know that why has he done that move? Why did it do that

move? It turned out that as the game built up and there's something rather different about chess

and go because chess gets simpler as the game goes on because pieces get taken off but go gets

more and more complex because more and more pieces get put on. As the game built up and more and

more pieces were put on the board. Territory was building up from the bottom right hand corner

and it turned out that AlphaGo's move at move 37 that black stone meant that it was it that

won that territory rather than Lisa Daul. It was an incredibly inspired move. It won AlphaGo the second

match that decision on move 37 to put the stone there and break the tradition of how humans thought

we should play the game. And for me this was really exciting because I believe this is an example

of what we should call a creative act by artificial intelligence. I spent some time on a committee

at the Royal Society over the last few years. We've been looking at the impact that machine learning

is having on society over the next kind of 10 years. Demis was on the committee and there was also

a philosopher Margaret Bowden and I talked to her quite a bit about the idea of creativity. She's

been very interested in what she calls these tin cans can do computers and the idea of whether they

could be creative. And she had a very nice working definition of what we should call creative. I'm not

sure it's the best one and we can argue there's lots of philosophical debate over the idea of what

we mean by creativity but I think this is going to be quite a useful working definition as we go

forward tonight. So creativity is something which should be new. Well computers can make new things quite

easily. We can objectively judge whether something is new but you have two other qualities.

It should also have an element of surprise. Now that's a little bit more subjective and also

value. That's also quite subjective. So a computer of it's going to be creative as something

that we as humans believe is creative. It's going to learn to know what we think is surprising and

has value. What do you think is so nice about games? It's of course you can judge this kind of

these qualities quite quickly. Surprise, yes, the commentators all went aw, it's made a mistake.

Value, yes, this move one alpha go the game. And so what I've been interested in is to look at

if it can be creative in this very close environment of a game where else can it be creative? Can

it be creative in mathematics? Actually I sat next to Demis and I joked to him we just become both

FRSs and I said oh could you get alpha go to become an FRS? And part of the story of my book is

about Demis said yeah we're already on the case. So they're already a deep mind looking at big

making a creative AI mathematician. So what I think is exciting about this new AI that is appearing

is that in the case of alpha go in not only played the game at a high level but it taught us how

to play the game in a new way. We thought we'd reach a kind of peak of a playing with these

rules that we had about playing in the one to fourth row in. There was an optimal way to play

the game. What alpha go has shown us in these games is that although we thought we were at the peak

of performance in playing this game, actually this was only what we met in a local maximum. Actually

this was like Snowdonia and there was actually a much higher mountain and Everest a new way to play

the game. But alpha go and experimented by taking risks on down this kind of adaptive valley and

found a much better way to play. And alpha go has now taught us a new way to play this game, new

strategies that are helping us to play the game at a much higher level. And so the journey of this

book which is called the Creativity Code is to look at well this AI that it's appearing that

seems to be able to learn about through its interaction with the kind of digital world around it

could it perhaps be creative in other realms not just creative in a game? In fact one of the first

people to think of the idea of code was already suggesting that code might be able to do things

of an artistic nature. So we celebrate Ada Lovelace day, Ada Lovelace every year, Ada Lovelace was

taken by her mother to see a Babaji's analytic engine. Her mother used to like to expose her to lots

different ideas, scientific ideas. And when she saw this machine she already began to realize that

this could do more than just the long division or the multiplication that it could do something

a little bit more exciting. And she started to write down code to make the analytic engine do

interesting things. And that's why we sort of celebrate Ada Lovelace the notes that she wrote

for a paper about the analytic engine. We regard as the first idea of code to make machines do

interesting things. Already then she was thinking about the fact that this could do maybe things

which are a little bit more interesting than just sort of scientific calculations. She wrote,

the engine might compose elaborate and scientific pieces of music of any degree of complexity

or extent. So she's already think about music, a place of course which has quite a lot of connection

with mathematics, the idea of patterns, getting the machine to kind of run out patterns,

perhaps it could make music and we'll come to that a little bit later on the challenge

of whether AI can write music. But she offered a word of caution when she wrote this. She said,

it is desirable to guard against the possibility of exaggerated ideas that might arise as the

powers of the analytic engine. It has no pretensions, whatever, to originate anything. It can do

whatever we order it to perform. And I think that's what we always felt in the past is that

this kind of top-down coding. Well, it's the human that's telling the computer what to do. So

if the computer is being creative, that's because the human has been created and it is encoded

that in just a set of rules that the computer is just implementing. But I think something has changed.

Now the code is beginning to change, mutate as it interacts with, say, new artistic data,

that it's starting to become code that the original codeer doesn't quite know how it's performing.

So this machine learning is producing programs which now perhaps disconnect itself from the original

codeer. So here's the challenge. Can this new AI that's appearing, which seems to be moving on

from the original code written by the codeer, kind of put some distance between the code and the

the codeer? So you probably heard of the Turing test, kind of computer part itself off in an

interaction online as I landed. Would you be convinced that that was a human talking or is it just an

AI computer? So Turing put this down as quite a big challenge, can it process natural language

and respond real time? So here's a new challenge that has been offered connected to the artistic

called the Lovelace test. So the test is can a machine originate a creative work of art?

Such that the process is repeatable. So it shouldn't just be some sort of glitch in the hardware.

Somehow the code should know what it's doing. It shouldn't be some sort of randomness which is put

in there such that the code wouldn't be able to reproduce what it's done. But here's the challenge.

Yet the programmer, the person who wrote the code that has now learned and mutated,

is unable to explain actually how the algorithm produced its output. So this is the challenge that

I want to explore with you. How good has AI been in the last couple of years? And it's really the

story of the book in understanding what we regard as art and creativity and being able to produce

its own version of that. And I think, you know, I think we're quite happy that AI is going to be

driving our cars or maybe even be our doctors or little nervous about, you know, it suggests

37 in the game two of go and suggests you take a pill that looks incredibly dangerous. Do you do

that? Is it a mistake or is it incredibly insightful move? It's going to save your life. So I think

we're quite happy in certain realms. But I think the one thing that we regard as the uniquely

human is our own creativity, our artistic output. That's what it means to be human. We express it

in music and art, in poetry and novels. So if AI can get close to doing something that

we regard as a uniquely human, I think this is a very exciting moment. So how good is it?

Well, I think I made a program for the BBC, a horizon of about six years ago. It was a touring

anniversary about AI. And six years ago, I was pretty disappointed in the state of AI at that point.

And there was one hurdle that seemed AI seemed to be finding really difficult to achieve.

And that was vision recognition. Our brain is very good at taking in a huge onslaught of

information, you know, lots of different colors here, people, faces. But I'm able to integrate this

into a single story about seeing an audience that's come to the talk I'm giving. So vision was one of

the great hurdles for AI at the time. And this is one of the things that it's being able to do

with this idea of machine learning. So actually we're going to do, we've used a bit of machine

learning learning to do and little view experiments during this lecture. So we have a camera here.

So the point is that machine learning, it shows some images of cats and dogs and it has

the distinguish them and it gets it wrong to start with. But it starts to ask more and more questions

which helps it to get it more and more right. So we've got to, we're going to be doing some

tests here. So you'll see you've got some cards in front of you. You're going to be doing,

I'm going to give you some challenges, some AI art kind of touring tests, love lace tests. And you're

going to have to decide what you think is made by a human and what do you think is made,

doesn't have a soul and you know, that's an AI. So in order to do this, right. So we're going to

get you up on your display one and we want the HDMI. So that's interesting. So display two,

I'm going to have, oh, okay. That's interesting. It worked and they're rehearsal exactly. But when

you're doing AI and tech, it's always asking for trouble. So, okay, well do you think you can sort

that out? It's not a disaster, but okay. So this is how we trained our AI. So what we did

was we want, when you throw up your cards, we want the AI to be able to recognize what you're

putting up that it's a blue, which will be for human and a red robot. But it doesn't want to

get confused by a red jacket over there or a blue shirt. And so we had to train this AI on pictures.

So what we did was we just took our random node of pictures and then we put

the things that we were trying to get the AI to recognize. And so there'll be a training image. So

we had about 600 training images. And basically it would be told there are four red robots and three

human faces. And gradually it learns over those 600 faces to be able to distinguish these.

Such that now, when we show it in image, it can count quite quickly and effectively how many

robots and how many humans there are. So we're going to get to you. So this is really machine learning

in kind of action. So what I'm going to do is to offer you some challenges. So here you are, this is

um, okay. So, you've, uh, I sure have to improvise. Okay. Um, so I'm going to, that's fine,

because I'll just swap over between the HTML and that will bring you up. So that's fine. Okay.

So here you are. So this, this is a project that was done. So I'm going to start with visual art

because vision has been the place where, um, uh, AI creative AI has been very successful. So

you can probably recognize the artist here. This is Rembrandt. And there was a team in Holland

that decided to see whether they could get their AI vision recognition to sort of understand

Rembrandt, very particular style of painting. Rembrandt did quite a lot of portraits. So it was

quite a bit of data, not as much data as we use to actually train our AI. Um, but you know, in the,

in the region of 300, um, portraits. And, uh, of course, one of the things Rembrandt has is a very

special use of light. So it was going to learn how to, um, put light on a portrait. Um, uh,

very particular style of, uh, dress at that particular time. So one of these images is a Rembrandt.

The other one is the product of the artificial intelligence learning on Rembrandt's style

and producing a new Rembrandt. So the challenge for you is can you tell, um, which of these, um, is the

real Rembrandt? Now I'm going to just do a little, um, experiment first of all to see whether

our cards are working. So let's switch over to, um, to show you. So here you are. So I'd like you

to have all put up your blue cards. Um, so, uh, let's see whether, and you can see it's starting to

some of you a little bit edgy. I'm sorry, you'll just have to feel like, um, your vote does count.

Um, um, um, and, uh, okay, now turn them over to Red. Let's see, um,

and you'll see the bar at the top is recording. So we'll be able to test the proportion. So that's

100% red now. Turn back to blue. Um, and you'll see the bar shoot back to the, to the blue side.

Okay. So good. So it seems to be working. Um, so now, um, not to prejudice the thing. Um, so I'm

going to, let's go back. I'll just show you the pictures again. Um, so, uh, I'm going to ask you

about one of these. So let me, uh, flip my coin. Um, so, okay. So I'm going to ask you about the painting

on the left. I want you to vote. When did you think the painting on the left is done by a AI, or is

it done by Rembrandt? So if you think the painting on the left is by AI, I want you to show your

red faces to me. And if you think it was a human, then I want you to show your blue faces. And

okay. So here we're starting to get, um, so quite a lot of you voting red and we're seeing,

um, although some of you are also voting blue. So it's, it's edging over, um, more to red. I'm saying

probably about 60, 30. Um, okay. So good. Let's see whether, how good you were at this one. Um,

so we'll go back to here. So which one was the AI? Yes. In fact, so you were pretty good already.

So you can feel good about yourselves. As an audience at least, uh, um, you voted correctly. So

yes, the one on the left is in fact, um, uh, the, the AI Rembrandt. Um, now it's interesting that

not only did they do this as a 2D image, but you know, if you've seen a Rembrandt, it has a very,

his use of paint is very special. It's very kind of 3D effect. And they even went to the extent

of analyzing the, um, the kind of height of the paint. And so they 3D printed this, interestingly.

Um, and when they asked, uh, Rembrandt expert to come along and, uh, review their, their, uh,

result. And of course the Rembrandt expert was incredibly snooty and dismissive about the whole

project. But the only thing that he could find to criticize it about was, well, the use of paint

just 20 years earlier than the style of the actual portrait. Um, so they, they team felt actually

quite good that they'd managed to, um, just, you know, that, if that was all that was wrong with

it. Um, okay. So, but you might say, well, what's the point about another Rembrandt? We've got

wonderful Rembrandt. Why do we need any more Rembrandt? So, certainly, my favorite, um, uh,

arkritic, uh, Jonathan Jones in the Guardian. I love reading Jonathan Jones because he's always

totally dismissive about anything to do with AI. Um, this is what he wrote about this Rembrandt project.

What a horrible, tasteless, insensitive, and soulless gravity of all that is creative in human

nature when technology is used for things it never should be used for. But frankly, anyone who

wears that sort of shirt is not crazy. I really, not quite sure. I trust very much in, um, uh,

their, their critique. But, but to a certain extent, he has a point. You know, what is the point

about creating other Rembrandt? Well, I do think there is a point because, uh, the wonderful thing

about this AI is it's starting to recognize things that we as humans have missed in the data. Um,

so, uh, not so much in Rembrandt. I haven't seen anything new in size in Rembrandt,

but for example, something like Jackson Pollock, um, a kind of algorithm to make analysis of Jackson

Pollock has revealed that Pollock is doing something very special when he splatters paint around,

that he's creating a very special mathematical shape that we can actually analyze and,

and kind of, uh, judge a kind of, uh, the fractal degree-ness of this dimension of these, uh,

painting. So, AI is giving us new insights. There's a, what are we sorry I'll tell in the book

about, um, uh, the Netflix algorithm, um, that just took, uh, our likes and dislikes of films

and the numbers of the films. Didn't know anything about the films,

but just from our likes and dislikes, there's able to clump them together

into films of a similar sort of genre. So you can see, oh yeah, look, these are all comedy films,

these are all, um, thrillers, but every now and again, it would clump films together because of

our likes and dislikes, which is kind of expressing our, um, uh, common, uh, feelings for a film,

in ways that we didn't really have a name for that genre. It was almost as if the AI had

sported through our likes and dislikes, that there was a kind of way of clumping films together

that deserved a new name. It had sported a new sort of structure in the films that we hadn't

kind of named. So I think there is a point about looking backwards. But I think the most exciting

thing is looking forward. Can we get the AI to do new things, to, to break them all, to do

exciting new things? So here's your next challenge. Four of these paintings,

a done by a human. Four of these paintings are done by an AI. You have to now judge, judge,

which is which. Um, so we'll do the same since I flipped. We'll do the left hand one again.

So the left hand one is the one you're going to be voting on. Do you think the four paintings on

the left are by the human or by the AI? So let's, uh, turn it over to you. So your chance to vote.

Um, those four paintings on the left, um, are they, uh, okay. So now, oh, much, oh, look,

there's Brexit vote going on there. Uh, yeah. Uh, so you seem to be still going for the,

it's much less convinced by that. But there's, it's just edging a little bit over, um, to the,

oh, yeah, it's still changing. All right. But I think that's, you think that one's the AI one. Okay.

Let's, um, go back and see, um, what? Um, uh, uh, uh, I think it's an awesome supporter over there.

What was that? Um, okay. So which one was which, um, no, in fact, those were four blue, uh, so a few

people, yes, I knew that. Um, yeah. So in fact, so you find that one a bit more difficult. Um, uh,

it's interesting because in some ways, I would say that the four which are produced by AI have

much greater complexity to them. Um, and these were actually shown at Basel Art Fair a couple

of years ago. I think this is 2016 these were. Um, and nobody was told there was any AI involved

in this. They were just asked to give their feedback on the paintings. And, um, the feedback on the

AI ones, they were, people were much more emotionally engaged with the AI ones than they were with

the human ones. And then of course you see, when you then tell somebody, well, in fact, there was

created by a computer, it really upsets people I think that, um, you know, oh my gosh, I had an

but that's, you know, there's no emotional world going on inside that. I think this is really

interesting. The reaction one has to, to experiencing something then finding out it's done by AI.

And I think most of you, I mean, I also feel like I've been cheated somehow. Um, but if I tell you

a joke, um, you know, Alan's jokes were all actually made by an AI. Um, uh, and you laughed at them

all. Um, but then if I tell you now, well, no, no, you actually just, you know, ran the AI, uh, joke

app, um, does that invalidate your laughter? I don't think it does. But why I don't think we

should get too threatened by this is because, um, this AI is learning on our emotional world to

produce its next step. So it's not disconnected. It has got an emotional world. It's representing

our emotional world, but in a sort of new filter. Um, what's interesting about this project,

I think especially, is the way that these paintings were created because these four paintings

on the right were not created by one algorithm, but two algorithms almost working in competition

against each other. They almost made it into a game. Something called a creative or generative

sometimes adversarial network. So the first algorithm was, uh, tasked with creating the art.

And what it did was to learn on all of the art of the past and it learned, it became a kind of art

historian. It learned how to classify art in particular styles. It understood when something was

cubist art or pointless art, um, by doing a machine learning process, um, uh, on the images and

being told what was, uh, in which particular style. Then it was tasked with creating something that

didn't fit into any of those styles. So it was really trying to break the mold. It had to make

something that couldn't be classified, um, given the parameters that it had learned. But it was

also tasked with creating something that we as humans would recognize as art. Um, so it had already

learned from the all, all the art of last 1,500 years, what we regarded as art. And so it knew

kind of an upper limit of where how much it could push the idea. The second algorithm was tasked,

it was to discriminate around with them, was tasked with, um, either saying, look, that, I think

that's still stuck, you're still stuck in cubist art there. Um, or else we're saying you've gone way

too far and that isn't art at all. And it was the competition between these two that ultimately

led to these images. Now, I think that's what's exciting with this particular, uh, idea

is because very often algorithms can just turn out loads of things, but the challenge is choosing

which ones are interesting. So we had a second algorithm, which was doing some choosing and discriminating.

And this is very close, I think, to actually how humans work creatively. Um, his Paul Clay talking

about the act of creation already at the very beginning of the productive act shortly after

the initial motion to create occurs the first counter motion, the initial movement of receptivity.

This means the creator controls whether what he is produced so far is good. There's always

that, you know, you do something and then you just, it's that good. I'm not sure I'm going to throw it

away. I'll do something again. Um, his Paul Valerie, a French poet talking about the idea of these

two kind of mindsets. Um, it takes two to invent anything. The one makes up combinations,

the other one chooses. And I certainly find that as a mathematician that I have collaborators

around the world where we kind of play these two roles. So I have a collaborator in Germany

where I'm the kind of mad creator and he's the discriminator kind of knocking things down.

Um, while it's a collaborator in the Middle East that I have, um, he's the kind of mad creator

and I'm the discriminator in that case. And by doing that kind of, uh, combination that we actually

make progress, um, together. Okay. Uh, so we've done the visual world. What about the written word?

Um, the written word. We already heard how, um, uh, text kind of prediction, uh, produces some

strange effects. Um, the written word interestingly, uh, AI is having quite a lot of difficulty with.

But then again, it's one of the first things that AI was actually, um, interested in trying to do.

This is, um, the Manchester Universal Computer. After touring left, uh,

Bletchley Park, he went up to try and realize some of his ideas in Manchester. Um, and the team there

were rather perplexed when, uh, letters started appearing around the lab, which were kind of

love letters, um, written by the Manchester Universal Computer. Um, and they were sort of

perplexed by this until one of the team admitted that in fact he'd written a program for the

computer, which was a template and he was using, um, a random number generator that, um, touring

had just created for the computer, which was randomly filling in the template with words,

uh, amorous words. Um, so after a while you'd spot the template, um, it's not very good.

But poetry is somewhere where AI has been quite successful. Um, I think partly because it's

again, nice closed form, it's not asking, uh, sort of too much large scale structure. Um, um, uh,

also I think that, uh, actually, you know, as an audience, you bring a lot of your own creativity

when you work, look at a piece of art, don't you? I mean, I think that's a point, uh, an artist

leaves room for your own, uh, world to fill, uh, things with. There's certain ambiguity to things.

And so, you know, poetry has a slight, no-mit quality that I think especially is something that

you bring a lot of your creativity to when you read it. Um, so here are your challenges now.

I've got some poems for you. I want you to vote whether you think these poems are,

gosh, this guy's already gone, oh, god, poetry now. Um, uh, so whether these poems are by AI

or are they by humans? So bought or not? Okay. So, um, here's your first poem. So, uh,

I'll read you the poem and then we'll go to, um, uh, to see what you think about it. Um,

I read it all. So, mortal, my mate, bearing my rocker heart, warm beats with cold beats company.

Shall I earlier or you fail at our force and lie the ruins of rifles once a world of art?

Okay, let's stop there. So, um, do you think that is bought or not? So now you're just voting, uh,

read for robots, um, uh, blue for, uh, oh, god, so I think you can get that one positive.

Okay, um, so as a massive vote for read there, um, uh, a few thinking it's, uh, human and blue.

Okay, so that was your first challenge. So let's go back to, um, your next challenge. I won't

reveal them yet. I'll reveal, I've got three poems for you. Um, okay, so here's your next one.

This is quite different, even three. There are smallances of pesticide reaction of real time,

of packs of displaced exclusionary heart, hurt, of powerlessness, of magazine fired,

non dignified, as head fatty implied internalized violence. A frozen helplessness is,

off white, cholola, did, did, did, um, okay. Um, so, what do you think? Do you think that is AI

or do you think that's human? Okay, so let's over to you to vote. Um,

am I messing with you? Or, you know, it's clearly, uh, okay, so you think, yeah, that's, uh,

so you, that's quite a little, yeah, a little bit for human there, but still some people thinking,

okay, is this a double block? Because that's clearly code. Um, uh, okay, so I think you've got

one for human there. Um, okay, so, uh, all right, this is working. Okay, so, um, um, uh,

here's your next challenge, bot or not. Imagine now the dark smoke awakened to fly all these

years to another day, notions of tangled trees, the other side of water. I see it is already here,

sequences of a face, see this shared and old friends pass their dreams.

Bots or not? Okay, so over to you. You think they're all bots. Okay, so, um, you're going for

bot on that one. So, um, uh, so that's right. So you think they're all, um, uh, AI. Yeah,

that's the sort of thing I would do, isn't it? You know, um, okay, so let's see, so you've gone for

the, for the bot on that one. So again, uh, do you, oh, they went human, yes, that's right,

they, sorry, you're absolutely right. Yeah. Um, thank you for picking that one up. Yeah, so,

um, let's go back. Let me give you the answers then. So, um, the first one you, or pretty

convinced that was AI, parole, Gerald Manley Hopkins will be turning in his grave. What? You think

I'm from below? Um, I chose Gerald Manley Hopkins because I've never understood any poem that

Gerald Manley Hopkins has ever written. Um, so didn't sniff that one out. Uh, second one,

yes, you did sniff out that, uh, okay, that's, that's too much like code. It can't be an AI. Um,

it's actually a young Australian poet called Mesbury's and she's very interested in this kind of

interplay between computer code having its own kind of poetry and rhythm to it, um, yet having

meaning that, uh, might have something to say to us. And so she's very interested in this kind of

weird interface between the two. So actually we've had two humans. So that leaves the last one

is that the human did I really mess with you. Uh, no, the last one you sniffed out actually

was the only one that made any sense at all. Um, was in fact the only one created by an AI. Um,

this is actually, uh, Ray Kurzwell created something called the cybernetic poet. Um, and this is a

machine learning, uh, process where he took poems of Yates, Keats, Elliot, um, and then the, uh,

the, uh, bot was kind of, uh, tasked with creating something, which was kind of fusion of Yates

and, uh, Elliot, for example. So, uh, of course Ray Kurzwell is, uh, one of the, uh, people who

talking about the idea of the singularity, the moment when computers might actually, um, be more

intelligent than humans and so that moment the singularity. Um, so poetry actually is not doing too

bad. Um, it's a kind of longer, uh, uh, uh, scale writing that AI is having real difficulty out.

So it can generate quite interesting sort of text, uh, generation or sort of small scale. Um,

but it's kind of the idea of writing a novel is still way beyond it. Although there have been

some attempts to write novels. So, um, there was a very interesting case, uh, uh, by, uh, a team called

Botnik. Um, they decided, um, they're big Harry Potter fans and they decided they were very disappointed

there were only seven volumes of Harry Potter and they wanted an eight. They wanted to know what

happens next. Um, so what they did was they got the machine learning to take all of J.K. Rowling's

writing, uh, learned kind of ideas that she's interested in, her style of writing. Um, and decided

they would create an algorithm to create an eighth book. Um, so here is the beginning of this

eighth book. I actually love the title. This is called Harry Potter and the portrait of what looked

like a large pile of ash. I'd read that. I'd read that. Um, so that it starts off pretty well. Uh,

magic. It was something that Harry Potter thought was very good. Um, so good. It's already picked

up that these books are about magic. You know, not bad. Um, leathery sheets of rain, lash

that Harry's ghost, uh, is he what leathery sheets of rain? I think that's a beautiful image.

Leathery sheets of rain. Uh, as he walked across the grounds towards the castle,

after this it began to lose the plot a little bit. Um, Ron was standing there and doing a kind of

frenzied tap dance. He saw Harry and immediately began to eat Hermione's family.

Well, that's good. Ron's wrong shirt was just as bad. It's Ron himself. Um, so AI is having,

it's very good at kind of local generation of things, which have some sort of meaning,

but it doesn't have very good sense of a long term structure. Okay. What about music?

Lovely. Say to gave us this challenge of whether AI could produce music and music has full

of lots of patterns. Uh, a composer when you hear a piece of music on the radio, you can probably

very quickly pick out what the composer is because they have particular styles, um,

particular sort of sound world that they have. Can the AI learn that and be able to produce

something at a particularly good level to, to replicate or do something new? Um, so AI always starts

on Bach. Bach is where AI always starts because Bach has a lot of algorithms at work. If you do,

look at something like the musical offerings, uh, Bach wrote the musical offerings, these little pieces

he wrote for, um, Duke Ferdinand as a, as a kind of puzzle to solve. There was a little algorithm

you had to show which you had to expand and, and see what the music actually meant. So,

so actually Bach is a very good place actually for AI to start. And some of you may remember a

few weeks ago, um, the Google Doodle, um, did you have a go on the Google Doodle? It was celebrating

Bach's birthday. And this Google Doodle, you could put in, um, a line of music and then it would

harmonize the other three voices. So the machine learning to take, this is based on, uh,

something called magenta. Um, it had taken all the corals that, um, Bach had written.

Corals are very good because they're kind of, they don't change, uh, key very often. They're

very sort of boring. Um, and, uh, but it can learn a lot about the way that a corale. And a

corale is filling in the harmonies a bit, like filling in a Sudoku. You've got to just learn the rules.

So this would actually, um, the, the little Google Doodle would fill in, uh, with your tune.

So actually put in the, uh, a bit of the musical offerings. So this is my, uh, attempt on it,

because the musical offering is actually quite a difficult challenge. It doesn't seem to have any key

to it at all. Um, and when I played it back, it was really rubbish actually. But what was very nice

about the Google Doodle is you could then say, I thought that was rubbish. And the thing would learn.

And actually say, okay, that harmonizing didn't work. And so you were training the algorithm as

you gave your feedback. And if you thought it was good, it would, it would kind of, uh, uh,

re-parameterizes, okay, I'll do more of that sort of harmonies. So very nice. So what we've done,

and I've got a, um, I actually set up a center in the Royal Northern College of Music with a

composer, Emily Howard. Um, it's a called PRISM, which stands to practice some research in science

and music. And what we're here really interested in is this kind of interplay, um, between the

scientific and mathematical world and the world of composition. Um, so we've done various projects

we wrote a string quartet representing mathematical proofs together. Um, but I've now got a,

I've now got a PhD student who's a composer, which is really exciting. Rob Laidlow, um, uh,

actually had his, uh, quartets, uh, premiered last week at the Wigmore Hall. Um, he was going to be here

tonight, but I'm fortunate he's, um, ill, uh, which is a real shame. Um, but, uh, he's been working

alongside me and various other people, um, as she here in the university as well, um, to produce

a piece, uh, using some of the best software we have around at the moment to create a kind of

hybrid, uh, AI bar piece. And this is what we want to play you. Um, so the challenge here is, um,

so we've got a fantastic, uh, musicians in the department here and Kobe is going to come up and

play this piece, uh, uh, for you, which is, uh, so in the, we just been doing very straight AI, um,

human tests. This is a slightly different challenge. Um, you've now got a piece which sometimes

as AI and sometimes is Bach. And I'm not going to tell you how many times it moves between one

and the other. And what I want you to do is to see whether you can see where the joins are. Can

you tell, oh, gosh, no, that's horrible. That's gone AI or, oh, yeah, that's Bach. Um, so we're going

to run this twice. Um, the piece is about four minutes long. Um, and so we're going to record

your thoughts on this. And then hopefully if we can do this, um, I'm going to play this back to you,

oh, gosh, because now we don't have two screens. That's going to be quite okay. Um, okay. No, I,

I know what I can do. Okay. Yeah. I thought of an improvised way of doing this. I'm going to do it

very analog. Um, uh, so you'll see what my solution in a minute, but, um, okay. So, uh, so, um,

perhaps we can give a big round of applause to Kobe who's going to come and play this piece.

We have a page turner as well. Um, and so I just to set this thing off, can you all show your

blue faces? So we're going to just start you off all, all on blue. So the idea is, as soon as you

think that the music has gone into something which is not by Bach, you turn to the red face,

and if you think it's gone back to Bach, you move to the, the blue face. Okay. It's very simple.

Okay. Um, so now in order to sync these things, I'm going to have to count this down. So we're

going, hopefully we can show these two things again. So I'll, you ready? Ready? Okay. So I'll go for

three, two, one.

Oh, that was amazing. I think there are moments when Bach will be turning in his grave, uh,

as you'll see. Um, but it's interesting because I think when it was Bach, there was much more

confidence, uh, in your knowledge of that. You could feel it was right. But when it wasn't,

it was really kind of edgy and there were moments when his suddenly surge red was,

there were giveaway moments. So, um, so, uh, we're going to replay this so you can actually see

what, what your answers were. Um, this was actually, so what we did was we took, um, uh, one of

the English suites. So the fourth English suite. And what we did was to take these, um, bits out

of the, uh, piece of music and then ask the AI to fill in the gaps. And the machine learning, uh,

was, it's interesting because we actually use quite a simple piece of machine learning. It's called

Clara. It's developed now. It's developed by OpenAI, this team, um, which is, uh, trying to make AI

very open to the world, uh, something that Elon Musk has helped, uh, set up. Um, and, uh, so this

actually by, uh, piece of software written by Christine Payne, um, as part of the OpenAI project.

And it's, it's growing. So we're going to keep on working on this to, to make it even better with

something called, uh, news net. But what's interesting is that this is kind of predictive in the sense

of it hears what's happened up to date and then makes a decision about what will be next.

So, um, some of the AI is working very cleverly and working backwards as well. So, um, um, uh,

knowing where the piece is going, it can make some prediction. But this piece of, uh, software

does not have a long-term memory. And this is one of the challenges to create a piece of

music software that can actually know about what it's done in the past and exploit that in its

decisions, uh, as it goes forward. Um, one of the things I talked about with Kobe beforehand,

he could feel the AI very clearly because the AI is not embodied. So it has no, um, trouble with

getting really awkward fingerings. Whilst Bach was writing something that would fit very nicely

under the fingers. Um, so I think this is one of the challenges of AI very generally is the idea

of it not being embodied. And you could really feel that, um, that it just did, I mean, when I first

played through it, it was just like, whoa, this is just really gooey. Um, okay. So let's, um,

play that back. So, uh, what I was hoping to do is have, uh, things on one screen and the other,

but I think what I will do, um, so we're going to, yeah, yeah, exactly. So let's, uh, we'll get

the, um, the thing up here. So we're going to get poor. Kobe's got to play this again, um, the,

painful little bits of the, the bark is great fun. And what I will do analog wise, um, so I was going

to do this on the screen, but I will just show you as we're going along which bits are AI and which

are not. And so you'll be able to see what you voted. And I'll tell you what the answers are.

Okay. Analog. That's right. Um, okay. So actually, I, I sort of slightly, um, biased things

by asking you to put all up the human face. And I wanted to see how long it would take

until you actually spotted that the opening wasn't bark at all. It was, in fact, so it took quite

a long time. So we're going to start. So we're going to count this down so we get the sinking right,

hopefully. So there you all are ready. So, uh, four, three, two, one.

You pick, you pick that up quite quickly that that was, um, bark. I think you can hear that it's

got some sort of direction to it.

You, you, you, you, you, you, you, you.

You're there.

You, you, you.

Or a ball.

Oh, yeah.

Now you can beat it, it's gonna bawg.

Cute.

Cute.

Cute.

Cute.

Cute.

Cute.

Cute.

Cute.

And thank you very much, Katie.

Now one of the things I was very strict about with Rob and the composer was that it had to be AI and he was not allowed any chance to try and improve the AI.

So we were very strict about because very often when you look at projects and many in the book, although it's saying it's an involvement of AI, you can see there's a lot of human input that makes a much better story of you just set the AI and the human is in the bolder tool.

So it was very strict with this, that the portions that where AI had to be only AI.

And interestingly, we asked it just to fill in the last chord and it missed out one crucial note which made the whole thing resolved.

But you can see from that that actually it was pretty convincing and there were few given lay moments but actually quite hard to pick out which was the AI.

I think it was pretty good on what was bark, although there were a few horrific moments when you thought it was AI and bark will be poor turning in his grave.

So I think that again, what's the point of this?

Well, I think there is a point for creative artists and I think this is not about competition.

This is about collaboration.

This is a new tool to push our own human creativity.

We've always already seen that in the realm of art.

But in music, one of the most interesting stories I saw was the idea of an AI that had been trained to play jazz.

It's called the jazz continuity to something constructed in Sony labs in Paris by Francois Pache and his team.

And they got the AI to learn in a very similar way to a jazz musician learning what the probability is of the kind of next move after a certain sequence of notes.

And they then did a concert where people found it very difficult to tell when it was the human playing or when it was the AI.

But what struck me was the jazz musician who had this AI been trained on.

His response to hearing the AI play back to him.

And he said, this is Bernard Lubad, the system shows me ideas I could have developed.

But what have taken me years to actually develop?

It is years ahead of me, yet everything it plays is unquestionably me.

And I think this is what's exciting because I think that we as humans often end up behaving very much like machines.

We get stuck in our ways of thinking.

We just perform the same ideas over and over again, especially in creativity.

I know that in my own mathematics.

I try the same things over and over again.

And sometimes I need something to push me out of the way I've been thinking and see that my world of possibilities is much richer.

It was like Bernard Lubad was in a room, the spotlight was on him.

He didn't realize it was so much more to play within his sound world.

So the exciting thing for me is that this movement into an AI that might be creative.

It's not about a threat, it's about an opportunity.

It's about the fact that this thing could push us to behave less like machines and actually become more creative again as humans.

Thank you.

Thank you.