Sep 14, 2020

In this episode Dan and Lee explore areas such as music and art and really try to explore what makes humans creative and can AI replicate that?  Can computers make Art, Music or even (one of Lees favourites, Wine!).

See further notes here:

https://www.amazon.com/Creativity-Code-Art-Innovation-Age/dp/0674988132

Human and AI Art collaboration  https://www.christies.com/features/A-collaboration-between-two-artists-one-human-one-a-machine-9332-1.aspx

Music and AI - https://bernardmarr.com/default.asp?contentID=1833

______

________________________________________

TRANSCRIPT For this episode of The AI in Education Podcast
Series: 3
Episode: 9

This transcript was auto-generated. If you spot any important errors, do feel free to email the podcast hosts for corrections.

Hi Lee, welcome to the AI podcast. How are you?
I'm I'm well done. I'm well done. It's good to be back and good to be talking AI again.
Yeah, definitely. This is going to be a brave one today because we haven't planned it um as as well as we could. We I thought we'd make this one really openended and really have a bit of a discussion. Um it basically came up uh in a couple of areas this week for me where through work and through social media, there's been a couple of examples where people have been talking about creativity and bringing technology into that equation as well. So, for example, I caught one tweet this week where where a teacher had been using iMovie and said, "Wow, I've made these fantastic movies and it reminded me of the time when Movie Maker was around and I was a moderator and exam board in the UK and I had 300 of these movies to check and you know, they all looked amazing but they all had the same pattern to them. They were using the same template
with a comics a comic sand serif spinning logo on the front.
Yeah. Yeah. Exactly. Well, actually, you know, you know what really kicked it was the kicker that Microsoft put in the the Star Wars transition where the writing goes up the screen and like every single kid across the world seemed to have a presentation that started with the, you know, the Star Wars transition at the beginning. Um, so and it it was so long and you had to watch everyone and I was remember sitting there for 300 of these assignments watching the Star Wars transition and it kind of you know, you get gets on your nerves after about 2 hours. Um, I'm sure.
Yeah, exactly. And it goes really slowly obviously, so you could read it all. But getting to the getting to the nub here, you know, it's always fascinated me this barrier between human creativity and artificial intelligence and whether you know if we can create AI and in in an authentic way then can that AI be creative? And I know we can start exposing you some of the creativity elements. What is creativity? What is art? What does that actually mean? But but it does seem when we Look at it at the top level. It seems a bit polar opposite. So on one side you've got creativity which we all think and love about humans thinking themselves outside the box coming up with original ideas. Humans are the only things that can do something creative almost or or the animal kingdom can do things that are creative. But it seems to be something about that originality and original ideas and inventing things that that give us that creativity. And then on the other side we've got AI which is an algorithm that's generated by machines and it's almost like this hard element to it. Um, and how how how on earth can that contribute to creativity?
Look, yeah, I it's this is an interesting one and it's probably good that we do a podcast on creativity creatively by not having any plan and just kind of winging it.
But but you know, as you were saying that and I'm listening, I'm thinking to myself, yeah, he's right in some ways that you know, obviously humans are the very example of just that random thought. You know, we just do things because we love it or we feel it. And AI kind of doesn't. But I got an example for you. So I was um in the notes thinking about last uh our last podcast around a responsible AI. One of the stories that springs in my mind is the uh story of Lisa Doll who's the um who is the uh Korean I believe uh world champion for Alph Go uh for for Go the game go
and few years back four or five years back he you know we alph the Google program engine to do to play the game of Go, which is considered to be the most complex game uh board game. I
I've got a confession to make. I nearly bought uh Go based on your comment on this uh a couple of episodes ago, and I nearly bought it last weekend, and it's so expensive. Like,
look, it's Yes, it's Yeah, it is.
That's unreal.
But but but but to the point about creativity, and this is I think this is a really good example. Uh in the game that he was playing, it was the whole story, you know, played a number of games against Alph Go. Uh and you know some won some lost some but what was really interesting the second game he played 37th move of that game um and I just know the specifics because I remember reading about it the the game of go has thousands of moves it's as I said a huge complex game I don't even begin to understand it but on the move 37 the system which had been programmed by watching thousands of previous games of go humans play lots of games so all it knew all it had ever seen is humans playing go on move 37 It played a move that and I don't understand how this can happen. It played a move that had never been played before ever.
It did something that that and and Lisa Doll who you know this this world champion stepped away from the board. He just it I mean to to use a a a good ' 90s expletive it just blew his mind. He just looked at it and said, "I can't compute what has just happened. I've seen something happen that
I don't even know if it's a mistake, if it's real, if it's even possible. So in the context of what you're just asking, you know, can create AI be creative? Well, I think in that example, AI did something creative. It created something new that didn't exist before.
Yeah.
So maybe yes, they can. Let's just not let's not close the podcast there, but yes.
Yeah. And that's the thing that's always kind of really really pushed my boundaries and my thinking because if we are designing, you know, if AI is designed for machines to think like humans and then go beyond human capability and then also find like the insights and the detail humans can't especially when it comes to really large detail complex problems then they must be able to find insights and new ways to think about things and and actually think a little bit creatively. So it is an interesting area. So another element um that I was thinking about was when you you know you look at some of the tools that that we got I was using a a program the other day 3D paint and I don't know if you've played with that or paint 3D um and and what you can do there's a magic select tool on there and you can draw around. So, I do this this demo regularly where you put a um I got a uh the pyramids of Giza and it's all overlaid, you know, it's a it's a one straight flat JPEG, but it's the Sphinx on in the front of the the pyramid of Giza. So, it's all just one flat JPEG image. And you can do magic select. You can draw around the Sphinx. And then you can actually um cut the Sphinx out using AI. And I know lots of Apple uh not Apple tools, the the Adobe tools do this as well. In Photoshop, you use it uses AI to fill the background in. It looks at the mood and the shadows in the back end and you can pull that the Sphinx directly out of that picture and it paints the the pyramid in the back end and it's you know there is an element of creativity in there because the person is repurposing something and using technology to augment they couldn't do previously. So the so the um the the thing that I do with that when I demo for example is I show the the the JPEG image. Then I take the sphinx out and then I'm using 3D paint and I paste that on top of a 3D image of a a a child and then you stick that on top then turn it into mixed reality. So you're being really creative but AI has done most of the work there. You know, you you look at it and you go, "Wow, Dan's now got the Sphinx like standing there as a child in the middle of the room in an augmented mixed reality pose and it looks phenomenal with the technology has done all of that. So, who's being creative there? Is it me being creative? Is it me using those AI tools to be creative? Um, it's a it's a little bit of a blur. Well, and and if to to use that example, uh, when Michelangelo painted the Sistine Chapel, was he being creative or was the painting the paintbrush being creative? I think in your scenario, you would argue that, you know, obviously Michelangelo was being creative. He the idea came from his sprung from his head and in your case, you were being creative. You created the thing. So, I don't I would argue that the technology wasn't very creative in that scenario.
Oh, okay. Yeah. And and that that that kind of suppose fixes in with um or connects or collaborates or whatever the word would be to to kind of say Marcus Datai wrote a wrote a book called the creativity code and he suggested in that book that I I wrote the quote down for that. So I did do a little bit of prep because I it was one of the articles I looked at. Um and um and his quote was machine and human collaboration that produces exciting results, novel approaches and combinations that likely wouldn't develop if you were working alone. So he's talking about the when we do collaborate um that we can augment and be creative more than if we were just working on our own. So those tools you know I suppose like you like you're saying there when you look at the Da Vinci and and this has happened all the way along life really hasn't it with you know when technology comes along whether it's a calculator you know all the mathematicians are saying you're losing the art of maths you know you now got a calculator you need to write it you can't do maths in a calculator you know suppose ink pens as well. That was another one when the ballpoint pen was invented. You're going to lose the craftsmanship of writing the keyboards killing writing the and and you know all of that kind of element and this was with the Leonardo da Vinci example the paintbrush probably people moourned at one point and said the paintbrush is killing art you know what was before that I suppose so let's art at the minute I know you did some work with the artist in residence program
in Microsoft can you tell us a little bit about that
yeah Yeah, look, I think it's I'm actually quite interested by that comment you make on Marcus Duta, that that creativity that the machines and humans collaborating to create something that wouldn't have been done if you were working alone because I think that's an interesting position to take on what is art. Um, because I think that's, you know, if we think about it, we're we're obviously we want to talk a lot about does AI kill creativity? Is AI adding value to art? And I think we could, I think a probably throwaway comment would be yes. But it was interesting as you were talking I then went look at okay what actually is art and every definition I can find of art starts with almost this exact same phrase art is the expression of human creative skill art is a diverse range of human activities and you can almost see it's by its very definition art is explained as a humanistic thing only humans can do it
now I'm not saying that's true or not but it's interesting that our generally accepted definition of art is only something that a human can do uh per se Hey, now I don't say I don't agree with that and I actually quite like what you're saying there with the Marcus Dotto Q point that when you bring two things together and if we consider AI to be another thought process human if like in inverted commas
and then working together you create something that's in the same way that uh you know the Wakowskis created the matrix as a as the two of them together you know they both built it not not one of them but both of them so you can you know you need partners to come together so Yeah, I it always harks back to that dis distinction of what is art to you and what is art to me because obviously you know art is hugely subjective and therefore if you choose to say well art is subjective is AI's contribution to art subjective and do you see it as being art or do you just seeing it as being a really smart computer program or do you see it as being a really intelligent person that's made a computer do something in a particular way the I think it always has always interested me in that area. I used to when you know I was talking about Mark in students work there early and I obviously have this debate with one of my colleagues um Jed his name was he's fantastic hope he's listening to this podcast but he was a art advisor for for a local authority in in the UK and school inspector a brilliant guy and um you know I always had this conversation with him how do you mark art how you know how could you actually mark it because marking is usually an algorithm a rubric that you put against something But it's subjective like you said. So how do you mark it? And if you can mark something then therefore there's a formula
to knowing if it's good or bad. Yeah. And we always just have that debate and he was very much around the process and it was very much of it's not about the outcome. You don't necessarily look at the painting and go wow that outcome is you know yes there's skills in there but your rubric would be about the skills used about the process about the thoughts about the types of uh investigations you done and the design elements and the composition. So actually there was a formula to mark art. So then you I'd be thinking well if you can if you know the formula to mark if there's an algorithm to art this this is often something you see quite a lot these days. There's a lot of artists out there have their own style and it's quite easy to repeat. You know if you go to Africa for example there's people on the streets and they painting the same painting over and over and over and over and over and over again. And you can say well is that creat Is it not? Is it mass production? Um, but they've got a formula that people go, "Wow, I love that art. It would be great on my wall." You know, they're not pushing boundaries. They're not doing anything that makes you think. They're doing something that's
pretty that looks nice in in my But isn't isn't that the very essence of art is I like it. I value it. I see it as something that appeals to me. It it creates an emotional response inside of me that I particularly like.
I mean, I think the way the example you're just giving there is um you know it's a really it's it's a really good one in the sense that you know when we think about what is art and can AI make art to your teaching example if you can if you can if there's a formula to determine what is good and bad art from a scoring mechanism a rubric as you say then that's in theory that's data you could give to an AI system and say this is how we score art and the system the AI would be able to go Oh, well, based on that, the very best data in the world and and you again, it's a good example. You've seen The Simpsons. Um, there's an episode of The Simpsons where Homer Simpson builds a car.
And he builds this car based on kind of the logic of what are all the most amazing things that every car has that you would want. And of course, you know, you take big four-wheel drive wheels, you take the massive stereo system, you build the, you know, the color painting, you have the all the bits that you think, and you think, "Yeah, yeah, must have that." And then you put it all together, and what you end up with is this is god awful mess. But what nobody wants and that's probably the risk with art and AI is if you gave it all these rules that your friend colleague has around how to mark art,
you'd end up with either the most boring art because it would just normalize to what everyone likes
or the most awful piece of art because it would just be so randomized trying to fit every single every single niche.
So it's hard to say. Yeah.
Yeah. Absolutely. And and you know when we thinking about say the technology element to that then and and that element of that art in resident program at Microsoft and these other projects that we've got. Um and and and I was just thinking about it as well, you know, you're you're right. Um you know, it depends what that how that art's created. And you know, I know this isn't like technology only about our art here, but but I think if you got a chimpanzee example that we were talking about just before we went on air just to discuss some of these things, your example about chimpanzee and would you consider that to be art if a chimpanzeee was making art? You know, somebody who's been following chimpanzees and their behavior patterns and things and then you give that elephant or chimpanzeee or whatever it might be a brush and they paint something then is the thought process the same you know it's wow this is quite a fascinating topic right
well it is and that's you know we should get back to the artist and resident piece because you asked about that but they but I think that's a good example people have done that they've they've given you know elephants and chimpanzees and animals uh paint and art and they've created artwork and you know to you and me maybe it looks like a lot of paint splatter on a wall but to somebody that really understands the journey that the chimp went through to get there. Yes, to them they can see something completely different in the picture and and that's what make art so art so unique. But I think to the artist and resent program because I think the really interesting thing here is the artist and resident program is a Microsoft research program and it's deeply linked to technology and art. How do we make sure that or how do we explore ways that art and technology can intersect? And again it kind of shines a light on the fact that what is what really is art? So the An example for me that struck struck me because I've seen it a few times and we've talked about it in in events is something called project Florence. Um so we had a a lady called Helen Steiner who was in an artist in resident she's an artist uh in resident and the idea here was to explore the potential explore the possibility of what if plants could talk and if plants could talk could we understand them and then could we have a conversation with them?
Wow.
And could we you know and and you kind of listen to that and you go that's I mean that's Crazily big idea. What an amazing idea. What's it got to do with art? Well, the creation of a new language, the interpretation of a language, the ability to understand what a biological entity is saying. And it's not saying like we're talking now. It's saying in its uh colors, its flourish, its uh
you know the I don't know the in the things that it gives off, the chemicals that it gives off. Do we understand what it's saying? And so the whole experiment was about understanding could we understand the plant. Now, is that art? I don't know. Is it AI? Absolutely. But it's all about using AI to cognitively assess all these data points about the plant and things that you and I couldn't see. So, light that is emitted from the plant and spectrums of of analysis of sound and other things that the plant is making that we can't see and hear. Definitely using AI created something new which was a a sort of a biological language that we could follow. But I don't know if everyone will call that art. hard one to validate.
It is, isn't it? And you mentioned sound there earlier on. So, I think it's about time that we kind of thought about the music element because this one's fascinating to me as well as a as a kind of amateur musician and a bit of a luier in my part-time, you know, when I when I look at music and you know that is all about algorithms. It is mathematical is this very formulaic and that that's always interested me because we have seen lots of projects appearing at the minute. I know there was one by some with some of our tools in um London I believe it was where they had a hotel that was playing lobby music based on the weather outside. So, it's interpreting the the clouds and then actually playing different music every time based on based on an algorithm which is quite interesting. But I suppose because music is algorithmically connected and there's only certain amount of you know again with you a kind of point about art and having a particular formula you know there's lots of songs where the same chords there's only a finite way of putting those chords together I suppose there's an infinite way of exp question those but music is an interesting one as well and I suppose the other the other element from that in terms of not only creating music but in creativity you know I have this conversation I I go to guitar the Australian guitar school every Wednesday there's a plug for the website there but um I'll put it in the show notes but I make guitars with a with a group and um yeah it's really interesting because you know that it takes a year roughly to make a guitar handmade you know that way and I find that really a creative process. But if you think about all the famous musician uh or musical instrument makers, I suppose globally like Savvarius or Leo Fender and and all of those folks who've made uh guitars and instruments that are the kind of top end and they've been very creative when they've created that, but then they've mass-produced them, you know, in factories and then those designs are mass-produced again in other factories for cheaper, less quality products sometimes and less quality work. But essentially exactly the same specifications, you know, where where does the creativity lie there, you know, can can creativity creativity be repeated um built upon, augmented, you know, so there's two elements in music for me. I think AI there's definitely a place for AI and and how that can be used to augment and and I suppose we use it at the minute in the music scene where people are developing um tracks for example, drum tracks for songs. AI can and I'll pick up the beat that you're playing on the guitar and put a backing track to you. So, there's a lot of AI starting to appear in recording as well when we record this podcast.
I use AI and I use noise gates and algorithms to compress and make us try to sound uh uh at the same level because we're using different microphones. We we speak I'm near my microphone. You you got a headset mic on. So, AI is used in music in all kinds of different areas. So, I don't know if you got any thoughts around that music. There's a couple there's two things I probably go back to your first one about the sort of the music itself. Um, so I'm a I'm a piano player. I've been playing piano for years. And I think when you look at it and you go, okay, you know, how many how many how many distinct sounds are there on a piano? You know, there's eight notes in an octave and then there's the kind of u multiple variances of that. But ultimately there's only eight notes. But as you say, music is a series of algorithms, you know, ways in which you play those notes in particular order that leads to the ultim in creativity in that, you know, music always sounds different. Whoever's playing it there because of the human that has applied themselves to it. And I know that there's some interesting uh pieces of music in in history where
it's not so much the notes you play, but it's the sound in between the notes. And that's the gaps and the the pauses and the and the the the sort of the randomness that gets introduced into it that makes it unique. You know, why is Beethoven's fifth any different to a piece of AI generated music. I mean, they both use the same keys, notes, scales, stands, octaves, everything else,
but they sound there's a uniqueness to it. And it's again, it always comes back to me about that emotional experience of what you get from that sound or that moment. So, that I don't particularly enjoy listening to uh bluegrass music. It's not my thing, but some people deeply affiliate affiliate with it. I love the sound of a piano. Some people hate it, you know, and you know all these things.
Yeah.
So I think that the musical creation bit of it is an interesting one. I think AI
absolutely and the good thing about music is it's a bit like uh a piano or a guitar any kind of music. It's very logical. I mean there are there are probably in simple terms a basic set of rules around how music works.
You know in terms of the scales and then and the notes and then you know you recognize your thirds and your fifths and your descendings and your ascendings and your you know the kind of
formula. Right. Yeah.
But but there's a logic to it. And if you taught a computer everything that's logical about music, then in theory, it should be able to create beautiful pieces of music that follow the naturally assumed conventions of how what sounds, you know, upbeat, slow, melodic, you know, melancholic, all those things.
And maybe it could build something amazing. Maybe it would just rip off some piece of music because look at the data and go the most amazing piece of music ever is, you know, fifths or whatever.
So,
I think the I think music could I think creating music is something that AI could do whether you appreciate it as an art again interpretation.
Jumping around there that that second point about making the actual tools I think you know when when you know say making the instruments and that bleeds into another area like the architectural element. We go through those norms the socially accepted norms what a guitar looks like and some people do break boundaries. I was looking at a video yesterday on Instagram where somebody some guitarist wanted to reach the he wanted extra frets on his guitar and he asked Luia to take away the cutaway and make all of this take away off the guitar basically. So there are people that innovate out there but I suppose using AI can start to innovate and some things that have been created like I think one of the buildings at UTS and some of that engineering element that AI being able to push boundaries and mathematically model beyond the comprehension of of even our best engineers is is fantastic. Fantastic.
So yeah, I look it's I was thinking as you were talking about the the creating the building the guitar because I know you build guitars and that's a very personal experience. You build them and you create them and you craft it. It's a it's got more of you in it than perhaps you realize because it is an extension of your thoughts on that and and you as you example you give as a as a human and understanding the mechanics of guitar sound and music, you might just go, "Hey, crazy idea. Let's put 24 strings on it. make the fretboard this long and do that. And you can think that way because you've got no idea what the outcome will be,
but you can think about it and you might have an idea and a a process around it. Whereas if you said to a an AI system and said like here's here's how you build a guitar and said to it, "Now go build a guitar." You can be damn sure it probably just build a guitar like the one you know because that's what it's learned. I think it's that interesting area where where you get to AI and creativity is how if you taught the AI system to say here is what makes a guitar pleasurable. Here is what makes a guitar sound
something humans like to listen to and you get into the sort of tonality and the sound and the and you're not talking about the building but actually what it is to be a guitar which is again a hard thing to easy to say very hard to explain I don't think we could explain that to a to a computer
um you know it's like the Douglas Adams situation of you know asking the question of what is the answer to life universe and everything but if you don't know how to ask the question yeah
then you won't get the right answer And that's so you couldn't ask us an AI system to build a guitar in the same way that you might go and build one that is totally unique because I just don't know that AI has that creativity to it to to take that stand off.
But taking that though, you know, I think the the the thing that would be useful for me would be to augment my build process. So for example, if I could when I'm doing the soundboard to that guitar, the top part that kind of resonates potentially the bracing underneath is really important to do with the sound and things like that. Maybe every piece of wood is unique and it sounds differently. Maybe using AI to train AI to listen to that. You know, you tap the top and it sounds differently. And some of the amazing luers will listen to it and say, "Oh, this is a great piece of wood." But having AI to kind of tell you, "Hey, this is this wood is quite thin. You might need to brace it in a particular way." Maybe suggesting bracing patterns for that specific piece of wood because The problem is we don't break that many boundaries. You know, if you look at a violin, it looks the same as it always has kind of thing. Uh there's me saying I'm not a violin maker, but guitars are predominantly made the same way. They've got multiple types of bracing systems, but there's a formula because they take so long to make. Nobody's really experimented and gone, I wonder if we could get a better sound if we do this. And some people have um but AI could help with that.
Yeah, I think that that the good point there. What you the human brings the creativity, the what if, the the hey, I wonder if we started to use XY Z wood or we built this approach. And then you're right, AI can kind of look at the mechanics of that and go, well, this is going to work better than that because of this change in the pitch tone situation. And that's but again, I think in that scenario, you're still at the point where
the human is creating, the technology is accelerating that creativity, is helping that creativity journey. along and and and I think it's still I mean I'm I'm sort of I'm wavering. When we started the conversation I threw you the the Alph Go story. I was kind of like well look hands down there it is mic drop AI can be creative but now as you kind of go through it I still see it's the human piece that really drives it. Um like an example for me Dan and is is wine. And if I think about wine to me you know I have this argument constantly with my wife because I have too much wine. Um Um because I enjoy collecting wine and I enjoy not so much even the taste of the wine although frankly yeah enjoy that
but I sometimes it's looking at a particular bottle of wine because of the place that it's come from because of the wine maker because of the the the history of that particular blend or brand or the the the vintage that it is. You know there's all those unique elements to it that to me almost make a bottle of wine a piece of art represent ative of a a memory, a moment or a thought or an emotion in me that I go I I want to hold that bottle because it means something in the same way that you want to hold a guitar because it means something to
and and I suppose the like the interesting thing though and you you probably know through these case studies but we've quite a lot of uh work around AI and the brewing industry um because obviously
I suppose you know when you are brewing and it's the same as any crops and a a industry I suppose But when we're looking at wine and and some of the other brewing techniques, you know, that takes a lot of time and those people creating the wines um you know in the various lovely wine regions in Australia for example, you know, a lot of care is taken and they do it at scale and if they do stuff something up or leave sediment in the bottom of the barrel for longer than they should, they might become more acidic or, you know, there's so many factors involved. So getting AI and sensor technology to kind of support the brewing is is quite important, right? Well, look, so it's a really good example. Yes, that we have we've done stories. I mean, there are examples of the AI developed whiskey and where, you know, and wine making and whiskey making, beer brewing, there's a kind of a there's a process to it. There is a set of steps. There's ingredients and a process and you can do it in multiple ways and get slightly different results and, you know, kind of create new flavors if you like or new things. And so, the AI whiskey one, I forget it was some uh where was it? it was over there Swedishbased whis distillery working on this kind of idea where they fed it all of the recipes for whiskey and labeled the data so taught the system what was good what had been sold well what was winners of awards and so on and then said to it create us a new whiskey now in that example the AI has created a new whiskey that didn't exist before creativity yes uh does everybody like it maybe not I don't know I don't know what the outcome of that was but it but it created something knew but only based on pre-existing knowledge. It didn't go crazy and say, "Hey, what if?" Now, in the wine- making industry, they've got a good example. It's everybody would know this one. Everybody enjoys a nice sticky um you know, a good dessert wine. So, largely speaking, the derivative of the betrodus um uh grape varietal or betrodus seion typically what we might use, but betroditis is a rot that appears on the grape vines.
Now, betus is It's a it's a rot that causes the the grapes to accelerate their fermentation. And what happens is the sugars get really really strong in them and hence why you end up with this deep sweet sticky strong dessert wine that we all enjoy. You know, a good a good glass of noble wine or something like that.
But if you said to any AI system, this is how you make wine,
and then have it say to you, you know what you should do? You should find the ones that are rotting and let them ferment a bit longer because you'll get this amazing other wine that nobody's ever drunk before. It wouldn't happen.
Yeah. But but humans, wine makers, you know, and that looked at this and went,
"Hey, what if we could do that? Let's try it." And I think that to me is the spark that today AI might struggle with because it genuinely is really making decisions based on pre-nown norms
versus, you know, a human that might say, "Hey, what if we did this?" And then used AI to help make that,
right? I I was up the hunter. I was at pigs. This is like a
favorite of mine.
Yeah, it's fantastic. I was We We can have really large show notes here. They're going to show our alcoholic and musical taste to the uh wine maker there uh recently when I was doing a wine tasting. Yeah, he's really lovely guy. Yeah, that's right. Lovely guy. And and and he was talking about one of the wines he made just because he just didn't have a wine to fit in his breakfast one day. He was like, "I really need a wine that's sort of like orange juice but kind of red wine." And and then he kind of made something and he came up with this fantastic pig's juice wine or whatever it was and and it was it was fantastic and you know he just kind of
knocked it together and
you're right. It was it was it was no particular formula for it. He just wanted to to go with with his breakfast.
Steve is a funny guy. I've just I went to Pigs Peak many years back. I was with with Steve went out to the back into the wine m into where he had his wine sheds. Yeah. And he was back before I think they were pretty big. And he's like he said
I want I want you to try something because we bought a lot of wine from him. And he went into the back and he went into the barrels and he just took the pipet and took a few samples from a few things of percentages, poured it into a glass. And he said, "I haven't made this yet, but I'm I'm thinking about making this into a new blend, you know, there's a bit of ganache and a bit of this and the other." And that's the kind of, you know, that that sort of that creative just, hey, let's just try something. It might explode. It might be good. It might be awful. Uh, but Steve's a good example of that. But yeah, exactly that. It's it's, you know, create something that didn't exist before.
Yeah. No, absolutely. And and I suppose, you know, I know we've kind of talked about some of these creative things like the the the music and and art and things but just something which just another you know to to kind of end with I suppose writing and novels and articles and things there's a lot of AI pervasively involved there in several reasons one to help you write help people who are dyslexic so giving a lot more accessibility to um writing books and things whereas I suppose previously you know you needed a typewriter you needed to be kind of visually pretty pretty good you might have to have a specific uh English degree Whereas these days AI is allowing lots of people to access and write their own content. I suppose technology being able to publish that quickly and digitally as well. What used to be this was a really creative and elite process has kind of really been democratized now and you know looking some of the books that have been really popular over the last 10 years you know including Harry Potter books and things like that they've come from almost amateur novelists I suppose. So that's kind of an interesting element to where AI has supported the writing process.
Yeah, look, it's Yeah, I I'm look, I I there's no doubt in my mind if we think about this, the creativity of writing, the creativity of music, the creativity of art, all of the sort of visual uh and food as well and wine and all these things. There's no doubt in my mind that AI can create things when given enough information. No question about it. Can it create good things? Sometimes. Does it create artist things. I think for me I'm on the fence and I would say that largely it's always in the eye of the beholder as is art. And so I think that if you'd spent your life programming AI systems and then you suddenly saw a piece of AI that wrote uh concerto, you would think that is art. But if you were a historian of the musics and you believed in the sanctity of the of the human, you know, music writing experience, you might look at it and call it a trans.
But but is it just on that one? I said just one Is it a turing test or you know is it a Turing test for for art? Is it a Turing test for wine? Is it a Turing test for for food and novels and and and I suppose it does also on a serious note also make us think especially on the writing element if AI can write an article or start to you know give us a feed of the news for the day then bias can come into that as well. So there's a there's there's underlying in some of this there's an element of um the previous around that kind of black box of AI and knowing that if I'm going to buy something, I know where it's been created and how it's been created possibly.
Yeah. No, I think I think actually you're on to something there, Dan. I think the touring test for art for AI creativity is perhaps where we should
and you need to know, right? Because I just thought just this horrific thought that you know, you could create an art piece of art for your uh living room based on an algorithm based on a famous speech you like, for example. like you said mentioned it in Martin Luther King speech but then somebody could put anything into that and put the speech by somebody you completely disagree with and then put that and you put that on the wall and then you don't even realize so you need to know that process that's been created I suppose and and to bring bring things back and and kind of end this podcast there I suppose what's your kind of general feeling on how AI is going to take us in the future around art and creativity
I think would be ludicrous of me as a lowly me to try and think about what AI could do because as we've seen and we know the bounds are unbounded the possibilities are unknown and I've seen art you know AI helping us talk to plants create emotional buildings uh project ADA we run is another one of those emotional is those artists and residents about creating emotional building I I don't I think if I have to look at it right now I'd say I think AI is a tool to help creativity, but AI isn't an engine of creativity in of itself. But in the future, I will be love to be proven wrong.
Fantastic. That's a great way to end there. Thanks again.