Learn about how AI mimics human creativity by selecting and combining impression molecules and answers to the question “is there anything unique about human creativity that AI can't capture?”

Dr. Esther Fee Reinhardt is an expert in AI and sound. She studied computer science and musicology at the Julius-Maximilians-University in Würzburg.

She completed her doctorate at the International Audio Laboratories Erlangen. Her research focuses on timbre analysis in audio signals using feature-based approaches and optimized artificial neural networks.

As an improvisation coach and composer, she is intensively involved with human creativity and relates this to the current mechanisms underlying hypothetically creative AI systems.

[Automatic captions by Autotekst using OpenAI Whisper V3. May contain recognition errors.]

[SPEAKER_00]: Welcome to the podcast Artificial Art.

[SPEAKER_00]: My name is Steinar Jeffs.

[SPEAKER_00]: I'm a musician and a music teacher, and in this podcast I'll be interviewing guests about technology and creativity.

Music

[SPEAKER_00]: So, welcome to Ester Fie Reinhardt.

[SPEAKER_00]: You're an expert in AI and sound.

[SPEAKER_00]: You've studied computer science and musicology at Julius Maximilians University in Würzburg.

[SPEAKER_00]: and you have completed a doctorate at the international audio laboratories Erlangen.

[SPEAKER_00]: Your research focuses on timbre analysis in audio signals using feature-based approaches and optimized artificial neural networks.

[SPEAKER_00]: And as an improvisation coach and composer, you have intensively involved with human creativity and relates this to the current mechanisms underlying hypothetically creative AI systems.

[SPEAKER_00]: So my first question is, what are you doing these days?

[SPEAKER_02]: I'm here giving a podcast.

[SPEAKER_02]: Thank you very much for having me.

[SPEAKER_02]: Yeah, at the moment, I'm a freelancer in music and AI.

[SPEAKER_02]: And as I have been for many years, a composer and musician.

[SPEAKER_02]: I love that at the moment I have the time to really productively play around with AI and do my music projects, because the last years there was a lot of work at the university.

[SPEAKER_02]: I'm very happy to have real own projects at the moment going on.

[SPEAKER_00]: And those projects are a combination of art and science, you could say?

[SPEAKER_02]: Yes, so my last big project was at an AI workshop where I created a system with three other group members based on my synesthesia because

[SPEAKER_02]: yeah when I when I know that a song is in a certain key, then I really have this impression of colors and I thought it would be so great to enhance the immersion by switching from one modality to another and.

[SPEAKER_02]: draw or paint in certain colors which is then translated into music yeah and that was.

[SPEAKER_02]: my last big project.

[SPEAKER_02]: It's called Chromatone.

[SPEAKER_02]: You can find it on GitHub.

[SPEAKER_00]: Nice.

[SPEAKER_00]: You were saying you have synesthesia.

[SPEAKER_00]: For those who don't know, could you explain what that is?

[SPEAKER_02]: That means that, for example, when I know that a song is in E flat, so mostly I also have absolute hearing, but a song is in A flat, then I

[SPEAKER_02]: have this strong impression of orange, for example, or in G major, it's a light blue.

[SPEAKER_02]: And yeah, those are strongly connected, this hearing and feeling colors.

[SPEAKER_02]: But it's not only those two layers, but it's also vocals.

[SPEAKER_02]: So vocals, I really mean like, for example, an E.

[SPEAKER_02]: is is red and um is um is an a in um ah it's complicated okay but and sometimes and also with numbers it's really sometimes strange for example when i when i tell someone someone my phone number and then um instead of uh a four i say yellow no one can follow me any longer yeah for me it's the same but as soon as the issue is kind of uh

[SPEAKER_00]: when impressions get blended.

[SPEAKER_00]: So I've heard about people that could smell a color, for instance, or like taste music.

[SPEAKER_00]: But in your case, it gets blended with your absolute pitch as well.

[SPEAKER_00]: And you kind of feel like a color.

[SPEAKER_00]: No.

[SPEAKER_02]: No.

[SPEAKER_02]: No, it's really more a probabilistic absolute hearing that I have.

[SPEAKER_02]: So yeah, I guess it in high probability, right.

[SPEAKER_02]: But yeah, I would not like to try that on stage.

[SPEAKER_00]: Okay.

[SPEAKER_00]: That's interesting.

[SPEAKER_00]: So I want to hear a bit about your background and then to your doctorate, which is called Single Musical Instrument Recognition, a solved problem.

[SPEAKER_00]: Could you explain that in a way that a layman could understand it?

[SPEAKER_02]: I try my best.

[SPEAKER_02]: You hear a tone and my algorithm says,

[SPEAKER_02]: That's a clarinet.

[SPEAKER_02]: So that's what I was doing.

[SPEAKER_02]: So there was a single tone from a single instrument, and this should be classified into the 11 most important orchestra instruments.

[SPEAKER_02]: So it sounds like a really well-defined and easy task.

[SPEAKER_02]: And I also went to a conference once, and someone was saying, oh, single music instrument recognition, isn't it solved already?

[SPEAKER_02]: And this sentence really hit me hard because I was working on this program for, for years and, and I was sure that I also know the literature.

[SPEAKER_02]: And, uh, then I thought a lot about, uh, how people come to say such things.

[SPEAKER_02]: And, uh, then I started to do, uh, do a science critic or, or to, to think about a lot about, uh, how we, how we write paper.

[SPEAKER_02]: papers.

[SPEAKER_02]: And for example, if you have a classification problem, and you only write in your abstract that we solve this problem of, for example, single music instrument recognition, with a probability of 98%.

[SPEAKER_02]: Yeah, everyone thinks like, okay, that sounds like solved.

[SPEAKER_02]: But if you distinguish between

[SPEAKER_02]: three instruments, for example, like harp, drums, and double bass, or if you distinguish between 20 instruments, including viola and violin, this is sometimes not really reflected in an abstract, or you have to read until the data page on the first page, and people have

[SPEAKER_02]: so many things to read.

[SPEAKER_02]: And if you're not really in the field, you just see high accuracy, accuracy scores.

[SPEAKER_02]: And you think that this is everything is solved.

[SPEAKER_02]: And that's why I really would like to cry out to the science community, we need good scores that describe the complexity of test data.

[SPEAKER_02]: So of course, not only an audio

[SPEAKER_02]: But in all classification tasks, yeah, we should really invent a good score that describes how hard this classification problem really is.

[SPEAKER_02]: So that was also one part of my thesis.

[SPEAKER_02]: That's why this subtitle, a solved problem, was included.

[SPEAKER_00]: But you also then worked on solving the problem?

[SPEAKER_02]: Yes, yes.

[SPEAKER_02]: So I...

[SPEAKER_02]: did evolutionary programming, that means that the architecture of the neural networks.

[SPEAKER_02]: No.

[SPEAKER_02]: Viola from a violin is never easy.

[SPEAKER_02]: It's a very hard problem, to be honest.

[SPEAKER_02]: Yes.

[SPEAKER_02]: Yeah.

[SPEAKER_02]: Yeah.

[SPEAKER_02]: Yes, because if you just say humans, then you're not only able to classify better than humans.

[SPEAKER_00]: I'll rephrase that, better than the experts, expert humans.

[SPEAKER_00]: um yes i think so i think so that it's not particularly when it comes to as as you were um making as a an example at the beginning just one note uh i think a person would have trouble classifying anything if it's just one stable note i need to hear some uh yeah some playing

[SPEAKER_02]: Yeah, but often sounds are very characteristic within one note.

[SPEAKER_02]: For example, flute, there is a high amount of noise in it.

[SPEAKER_02]: But for example, the clarinet is missing the second and the fourth overtone.

[SPEAKER_02]: And this is something which is very easy to understand by a machine.

[SPEAKER_02]: But yeah, you have to listen a lot that as a human, you can distinguish this from other instruments.

[SPEAKER_02]: That's one of the rare examples where it's easier to see in some

[SPEAKER_02]: easy transformation, like the Fourier transform, but really not so easy to find out by our brain.

[SPEAKER_00]: What could such software be used to, you think?

[SPEAKER_02]: For automatic labeling of YouTube videos, for example.

[SPEAKER_02]: So that we could search the

[SPEAKER_02]: saxophone version of our favorite song in YouTube and that would be one use case for this.

[SPEAKER_00]: I've also heard some talk about the possibility of using some kind of AI classifier to determine which musicians play on which tunes.

[SPEAKER_00]: to enhance like credits in spotify for instance because often that's a thing that's lacking you only get the one artist but it doesn't say who played drums on it for instance maybe that's another work for such a classifier yeah interesting yeah

[SPEAKER_00]: So I'll go a bit further into AI's role in creativity now.

[SPEAKER_00]: And in your article, which is called a comparison of human creativity and generative music production, you discuss how AI can mimic human creativity by selecting and combining what you call impression molecules.

[SPEAKER_00]: Could you explain what that is?

[SPEAKER_00]: And do you think AI could ever truly replicate the depth of human creativity?

[SPEAKER_00]: Or is there something inherently unique about human creativity that AI can't capture?

[SPEAKER_02]: So, sorry, my first thought when I realized what AI is able to do was not that, oh my God,

[SPEAKER_02]: AI is so intelligent or so creative as we humans are but I thought that oh my god we are so stupid and so simple like computers are working because this is a really sad thought I think we can explain the main the biggest part of our creativity by

[SPEAKER_02]: By this we are watching something, we have an impression of something and then depending on our attention we are saving this only short or deeper or in connection as we said at the beginning with the fire and wire together.

[SPEAKER_02]: with other impressions.

[SPEAKER_02]: And then when we get creative, we remember those events.

[SPEAKER_02]: And we are saving this in different granularity.

[SPEAKER_02]: So for example,

[SPEAKER_02]: If we see a tree on a hill, then we're not only saving this picture, but also the warmth of the sun and the exact color.

[SPEAKER_02]: For example, the exact color.

[SPEAKER_02]: This is not only green, but if we have our attention on this leaf, then we are able in an amazing detail to save this exact color.

[SPEAKER_02]: And then when we are creative, we recombine this pool of all our former impressions, this can be the same for music, and depending on how fine this granularity is,

[SPEAKER_02]: how common it is to recombine this in a certain way.

[SPEAKER_02]: It is new or it's normal, it's the norm.

[SPEAKER_02]: So for example, if we take the same chord progression

[SPEAKER_02]: and the same sound as we already heard in many, many, many, many, many songs before, then we will create a great party hit, maybe.

[SPEAKER_02]: Yeah, this is sad enough.

[SPEAKER_02]: But for example, Stockhausen was not only combining

[SPEAKER_02]: instruments or tones but he was only using frequencies and i personally think he's the father of dubstep because he was using frequency ranges or single frequencies and noises yeah and even this highly creative thing can be explained by his former impressions but take it so

[SPEAKER_02]: so fine, such a fine granularity that you could even talk about like atoms.

[SPEAKER_02]: So it's no longer the molecules.

[SPEAKER_02]: He really took the atoms and recombined it.

[SPEAKER_02]: Yeah, so I'm sorry.

[SPEAKER_02]: The main difference, what's happening in our brain and in the bottleneck or the latent space of an autoencoder is that we choose from this pool depending on our emotion.

[SPEAKER_02]: and depending on our vision we have of a final result we want to achieve and if this is really matching with what we create then we have a good feeling so this is our fitness function is depending on our feeling if this is matching with with what we want and as

[SPEAKER_02]: I think all artists feel the same.

[SPEAKER_02]: This is the greatest feeling ever if you create something which is really matching with what you feel at this moment.

[SPEAKER_02]: And of course, in autoencoder, it's the mean.

[SPEAKER_02]: So the mean is depending on if it's going to be a good result or not.

[SPEAKER_00]: So let's see if I got this right.

[SPEAKER_00]: You're saying that we as creative beings, we take in impressions and we take in impressions of different kinds of granularity from like very coarse granularity, which might be taking a full chord progression from a tune.

[SPEAKER_00]: or with a really fine granularity like you were talking about impression molecules or perhaps even impression atoms or impression quarks even, which might be a single frequency.

[SPEAKER_00]: And then we can recombine these molecules into something new.

[SPEAKER_00]: And this is pretty similar to what a neural network also might do.

[SPEAKER_00]: But the difference between the machine and the human is that we can make a connection, I guess you could say, between the impression molecules and our emotional state.

[SPEAKER_00]: Is that about right?

[SPEAKER_00]: And I guess it seems like one is more creative if one uses fine granularity in contrast to coarse granularity.

[SPEAKER_02]: I think that creativity is not so much depending on the

[SPEAKER_02]: the novel team because children can be highly creative and absolutely free of fear.

[SPEAKER_02]: And the end result may be very, very standard.

[SPEAKER_02]: But I would say that this is still a very creative thing I would never take away from a child because I saw something like this before.

[SPEAKER_01]: Yeah.

[SPEAKER_02]: Yes.

[SPEAKER_02]: but if it is interesting and if it is as art interesting for someone else to perceive it this is depending on how new this idea is so you can't evaluate creativity based on the outcome then yeah yeah it's more about this reflection

[SPEAKER_02]: within the process, if this is matching to your own vision of so okay, I think creativity is this process of repeating to compare what you're doing at the moment and if it's matching with your vision of something that I think this, this I would take as a definition of being creative.

[SPEAKER_00]: It has to start with a vision then.

[SPEAKER_02]: The vision will come if you're not starting with the vision.

[SPEAKER_02]: So that's the best thing about ChatGPT and Co.

[SPEAKER_02]: You type something very small in and then you see something and then suddenly you know that's not what you want or that goes in some interesting direction.

[SPEAKER_02]: But you don't have to need the vision before because you will generate it in the first step if you take a pen or if you write something with ChatGPT.

[SPEAKER_00]: Nice.

[SPEAKER_00]: You mentioned there for a second about children not being afraid in their creative process and that's kind of liberating for creativity.

[SPEAKER_00]: And in your article, you also highlight that AI, unlike humans, create without fear of judgment.

[SPEAKER_00]: And do you think this fearlessness is an advantage or a limitation in the context of artistic creativity and how does it affect the final output, do you think?

[SPEAKER_02]: It's a huge benefit that machines have here.

[SPEAKER_02]: Yeah, it can't be overrated what we humans suffer when we create something.

[SPEAKER_02]: Yeah.

[SPEAKER_02]: Definitely.

[SPEAKER_02]: And in the end, the output is not better because you suffered at the beginning if you're ever going to be good enough or if it only has to be good enough for you.

[SPEAKER_02]: And the...

[SPEAKER_02]: If you manage to create something which is good enough for you, then it's not important that you also suffered what someone else will say about that.

[SPEAKER_02]: That decreases art.

[SPEAKER_00]: Often the greatest judgment comes from yourself, no?

[SPEAKER_02]: Yeah, we are hard enough on ourselves.

[SPEAKER_00]: A devil's advocate on that take of fearlessness might be that suffering is a part of the human condition which might also be a source of inspiration and also a source of great art.

[SPEAKER_00]: So if you don't suffer enough in the process might that make the art less valuable?

[SPEAKER_02]: You're absolutely right.

[SPEAKER_02]: So the ability to be afraid of dying, that our time is limited, that's, in my opinion, the main reason for all our feelings and the main reason why we're doing art and we are creative.

[SPEAKER_02]: So nothing has meaning without this fear.

[SPEAKER_00]: So one does not need the fear of judgment then because you always have the fear of death waiting around the corner, getting you inspired.

[SPEAKER_00]: Who's the author of that book again about Ernest Becker or something, is it?

[SPEAKER_00]: Which is about that concept of fear of death being the ultimate motivator of everything.

[SPEAKER_02]: Yeah, so I'm sure that this is a concept where when you think about that, I think it becomes pretty clear that this is the main point of everything, even of love.

[SPEAKER_02]: So a lot of people say that we have the fear of death and love as the two driving emotions.

[SPEAKER_02]: I think that it all comes from the fear of death.

[SPEAKER_00]: Yeah, I just had to check it out.

[SPEAKER_00]: And it's an author called Ernest Becker who wrote the book, The Denial of Death and the Practice of Dying, which is about exactly the thing you were talking about.

[SPEAKER_02]: I heard that there, unfortunately, I don't know the names, but there's some kind of hospice.

[SPEAKER_02]: So a place where you go to when you know that you're in your last month

[SPEAKER_02]: And there are a lot of art courses there.

[SPEAKER_02]: And this is so amazing that so many people, that you see that everyone has a creative potential.

[SPEAKER_02]: And for some, it only comes out when they know that they have nothing to lose.

[SPEAKER_02]: And then it's really deep.

[SPEAKER_02]: And yeah, so I really think that we all

[SPEAKER_02]: are artists but it's a lot about bravery how deep you want to dive into yourself and to show yourself to the world yeah and when you're when you have nothing left to lose then yeah nothing holds you back and you think AI could could help in this endeavor so that you don't have to wait until your deathbed before you create something

[SPEAKER_02]: Yes, in general, yes.

[SPEAKER_02]: Because you always want to write a song about blah, blah, blah.

[SPEAKER_02]: And then you never dare to start because you think that it's never going to be good.

[SPEAKER_02]: And maybe now you do it with one sentence, with suno, and then you get something.

[SPEAKER_02]: And then suddenly you feel like, OK, now I have something.

[SPEAKER_02]: I don't want it to be like this and like this.

[SPEAKER_02]: And then you start really to create.

[SPEAKER_02]: But this fear of the white paper can go away with this role of AI, where AI is a stupid, creative tool.

[SPEAKER_02]: And why I say stupid is I think AI can have at least three different roles.

[SPEAKER_02]: And the first one is this.

[SPEAKER_02]: I only write something very small.

[SPEAKER_02]: The human input is very, very, very small.

[SPEAKER_02]: And I create something, in the most cases, amazing.

[SPEAKER_02]: I believe that this will change in the next years all the music which is not used for a highly emotional context.

[SPEAKER_02]: The second role is

[SPEAKER_02]: to use AI as a tool, as an artist.

[SPEAKER_02]: So for example, last week, I put the sound of the bells from the Munich Town Hall into an AI, and text prompted a lot.

[SPEAKER_02]: And now I take snippets of the sound, and I'm combining this.

[SPEAKER_02]: So I would say that here, AI is really only used as my tool.

[SPEAKER_02]: And the last role is this really amazing ability of AI to translate from one modality to another, like I told you before with my chromatone application.

[SPEAKER_02]: And if you are afraid of the blank page, then also the first role I was talking about, this

[SPEAKER_02]: Please generate me something where I can say that I want it like this, or definitely you want to have something else.

[SPEAKER_02]: Here also this rule is beneficial.

[SPEAKER_00]: So you can make some new cutting edge weird art.

[SPEAKER_00]: You can do change of modalities, and you can make more entertaining music more easily.

[SPEAKER_00]: One thing that's often mentioned in conjunction with AI tools and composition and artistic expression is the challenge of authenticity.

[SPEAKER_00]: And that could be a limiting factor both ways because composers could be limited by the fact that they are afraid of being perceived as fake, that people won't be interested in listening to the music if they perceive it as made by AI.

[SPEAKER_00]: And I've seen this now multiple times, especially in the visual space.

[SPEAKER_00]: For instance, a couple of weeks ago, a band or an artist named Direct Trucks released a poster for an upcoming concert or an upcoming concert series.

[SPEAKER_00]: on Instagram.

[SPEAKER_00]: And then one of the comments said, this poster is obviously made by AI.

[SPEAKER_00]: You guys are ruining the graphic industry.

[SPEAKER_00]: And there were loads of comments of this kind.

[SPEAKER_00]: And then the thing that ultimately happened was Derek Trucks took the post down, removed the poster.

[SPEAKER_00]: and said that they didn't know that it was made by AI, but they had investigated further and found out that it was.

[SPEAKER_00]: And yeah, they didn't want to support that kind of thing.

[SPEAKER_00]: So yeah, it just points to how important many people find authenticity and the human touch.

[SPEAKER_02]: last week I was showing my music to a friend and then he asked about a song which was created absolutely without AI that this one must be the one with AI and I said no that's the one and I really felt how important it is for me that he knows that this is the song where really everything was handmade and

[SPEAKER_02]: Yeah, it was an interesting reflection how important it is.

[SPEAKER_02]: I think it's not about if I used an AI tool or not.

[SPEAKER_02]: If I still spend 50 hours on a few minutes, then I know that there is so much from me in this music that I can still proudly say that I put my heart in it.

[SPEAKER_02]: So I'm also fine with that.

[SPEAKER_02]: But if I did everything without AI, I also want the people to know that it's not done.

[SPEAKER_00]: Maybe it's like when you cook dinner for someone and they say, oh, it tastes great.

[SPEAKER_00]: And then you like proudly say, yeah, I made everything from scratch.

[SPEAKER_01]: Yeah, that's a good comparison.

[SPEAKER_00]: I also did hear the opposite view on authenticity the other day in an interview with Databots, the creator of Databots, who said that in their world, it was important that things that was portrayed as AI music actually was 100% made by AI.

[SPEAKER_00]: So if they listened to some AI music and they didn't have like a bit of noise or kind of AI artifacts in it, then it wasn't true AI music.

[SPEAKER_00]: So that wasn't authentic either.

[SPEAKER_02]: Yeah.

[SPEAKER_02]: But I think that with the increasing quality of Suno & Co.,

[SPEAKER_02]: the artifacts will be fewer and fewer.

[SPEAKER_00]: So then it perhaps comes down to how novel it sounds, I guess.

[SPEAKER_02]: On the other hand, I also created really amazing novel kind of music with AI.

[SPEAKER_02]: It's really creepy, to be honest.

[SPEAKER_02]: So my first Suno experiences was really like... I really could have cried because it was so good.

[SPEAKER_02]: Yes.

[SPEAKER_02]: I think that really everything with...

[SPEAKER_02]: I call it the man value.

[SPEAKER_02]: I hope that's not too anti-feministic.

[SPEAKER_02]: But I always say that I'm the mother of the man value.

[SPEAKER_02]: And I only found out that this is a variation.

[SPEAKER_02]: And the long version is minimal emotion needed.

[SPEAKER_02]: So, for example, if you have an elevator and you want to buy music for your elevator so that your employees don't have to have the same...

[SPEAKER_02]: conversation with their colleagues every day, standing there for 20 seconds, or to reduce the probability for claustrophobia, then the music you need here has low minimal emotion needed.

[SPEAKER_02]: So it has a low man value.

[SPEAKER_02]: And all this kind of music with a low man value

[SPEAKER_02]: will be, of course, replaced by AI music in the future.

[SPEAKER_02]: Also, when you go to an Italian restaurant and you own this restaurant, you want something which sounds a little bit like the standard Italo-pop songs.

[SPEAKER_02]: You don't need the original because everyone is no longer even happy about hearing the original because we heard it like really 5,000 times, always the same songs.

[SPEAKER_02]: But for example, on the other side of this scale, there are live concert experiences.

[SPEAKER_02]: For example, if you saw Joe Cocker on stage or on video and you see him almost dying on his passion when he's performing, then you know that it's not about the acoustic quality and it's really about experiencing the emotion of another human.

[SPEAKER_02]: who's giving everything.

[SPEAKER_02]: And so with this extremely high man value, such experiences can't be replaced by AI.

[SPEAKER_00]: Yeah, so the emotional value and the nature of seeing someone as well.

[SPEAKER_00]: So the combination of the two is really important.

[SPEAKER_00]: Regarding your argument when it comes to restaurant music or elevator music, that one has heard a song 5,000 times, so now it would be better to have just...

[SPEAKER_00]: a generated sound-alike version.

[SPEAKER_00]: The counter argument to that might be that humans really enjoy listening to the same thing 5,000 times.

[SPEAKER_02]: Yes.

[SPEAKER_02]: Yeah, you're absolutely right.

[SPEAKER_02]: But if you own this restaurant and you have to pay the GEMA, then... And it's not about that I suggest that this is a good idea.

[SPEAKER_02]: Go past the artists.

[SPEAKER_02]: No.

[SPEAKER_02]: But...

[SPEAKER_02]: Yeah, I think it's going to happen.

[SPEAKER_00]: It all depends what is the cheaper alternative then, if it's generated music or copyrighted music.

[SPEAKER_02]: Yeah, but I think it was always a little bit unfair because we have so many good artists and this, as you said, this principle of music being liked, if it's very similar or the same, which we already heard before, this always excluded a lot of...

[SPEAKER_00]: interesting and new artists yeah that's true and it's uh it's often said that your musical taste gets developed and kind of set in stone and you're in your teens in your teenage years and then it doesn't really change or at least for the the average human they generally listen to the same music as they did when they were like 15 to 19.

[SPEAKER_00]: um and i guess that's yeah we're kind of nostalgic a nostalgic race i guess so it might not i mean if you're at an italian restaurant listening to sound like generative music it might not evoke the same kind of nostalgia as hearing the exact song you heard there for 20 years earlier i don't know we'll have to see

[SPEAKER_00]: But on the question on whether AI can create art or not, what's your perspective on this?

[SPEAKER_02]: I think we live now in an absolutely oversupply time of everything.

[SPEAKER_02]: Also of possibilities to spend your time on art

[SPEAKER_02]: on experiences and so on.

[SPEAKER_02]: And so you have to carefully choose where you spend your energy and your time.

[SPEAKER_02]: And when I know that some picture is generated by AI and the AI hype is over, then I don't see a reason why I should use my life to watch something which is just generated by pressing a button.

[SPEAKER_00]: um yeah so that's that's my my opinion i just have to quote you and your article uh regarding that statement because you say what is the reason to devote oneself to something that was created without devotion that's a great quote yeah

[SPEAKER_02]: Yeah, that's really what I feel, because I once listened to, I think, a 20-second snippet, maybe 30-second snippet of music, and it got me.

[SPEAKER_02]: So I was really emotionally pushed.

[SPEAKER_02]: And then I found out it was AI.

[SPEAKER_02]: And then I really felt like...

[SPEAKER_02]: It really felt so bad.

[SPEAKER_02]: Yeah.

[SPEAKER_02]: And then I recognized that, why is this important for me?

[SPEAKER_02]: And the sentence came up, I don't want to open myself to something which was made without love or without fear or without emotion and energy.

[SPEAKER_00]: So with that in mind, one needs the devotion or touch of a human to create art that is interesting to other humans?

[SPEAKER_02]: Yes.

[SPEAKER_02]: I think we are social creatures from the start.

[SPEAKER_02]: And this wish to connect with other humans and to find out what they feel and think.

[SPEAKER_02]: This is a main concept.

[SPEAKER_02]: So this is one really basic concept.

[SPEAKER_02]: And that's the reason why we are interested in art done by someone else.

[SPEAKER_00]: um also in your article you mentioned that ai lacks the shared enthusiasm for a shared visions that human collaborators might have how do you overcome this lack of emotional synchronicity when working with ai in your creative process um an ai co-creator is like a mirror which is also very interesting but

[SPEAKER_02]: A human co-creator can really not only add, but really, really, really enhance the creative energy if you have a shared vision of something.

[SPEAKER_02]: Yeah, this is more than only this mirror.

[SPEAKER_00]: What kind of tools are you currently excited about and what are you using for your own art?

[SPEAKER_02]: ChatGPT is a creative motivator and Suno where I take snippets and sometimes only this role of prompting a sentence and impressing people

[SPEAKER_02]: And they are so shocked and I understand completely why they are so shocked.

[SPEAKER_00]: Yeah.

[SPEAKER_00]: When it comes to using Suno for your own work, do you like make a text prompt, get a result and then use that as inspiration to make something or just prompting and then that's it?

[SPEAKER_02]: I would really like to label my songs with...

[SPEAKER_02]: complete AI generated or AI as a tool or only AI snippets and no AI at all.

[SPEAKER_02]: So I should really create some something like labels where I add this to songs.

[SPEAKER_00]: What's your what's your dream tool?

[SPEAKER_00]: Your AI dream tool?

[SPEAKER_02]: So OK, I'm telling you.

[SPEAKER_02]: I would love to create a synthesizer where you have a VR set, and you're standing in a complete, totally dark, black room.

[SPEAKER_02]: And you have a glowing globe in front of you.

[SPEAKER_02]: And this is your sound.

[SPEAKER_02]: And then you're changing the shape, and you're changing the color.

[SPEAKER_02]: And you're changing the loudness by making it bigger or smaller and the intensity.

[SPEAKER_02]: And yeah, this is your synthesizer.

[SPEAKER_02]: And the attributes you're changing by changing the shape of this globe.

[SPEAKER_02]: And then you put different globes, glowing globes, into this huge room.

[SPEAKER_02]: And then you can play this as an instrument in this virtual immersive environment.

[SPEAKER_02]: room that would be my yeah i wish for this this ai too sounds cool will you make it good you want to build it yeah maybe um i think i think it's a it's a bigger project but i'm dreaming of this over 10 years um yeah looks like it's now it's uh

[SPEAKER_00]: achievable to to really to program this yeah uh one question back to something we talked about earlier which uh i thought about but forgot about asking um when it comes to what you described as the man value when it comes to level of emotional emotional investment you perhaps could say

[SPEAKER_00]: that it might be even more important in the arts to have art that has high man value or high emotional connection.

[SPEAKER_00]: And this also reminded me of a statement from Sam Altman, the CEO of OpenAI, who said something of the sorts that he reckons most jobs that don't require a high emotional connection between two people, those jobs will likely be gone.

[SPEAKER_00]: And you also...

[SPEAKER_00]: You also designate kind of a gender category to those values of emotion in terms of the man value thing.

[SPEAKER_00]: And also a common conception that women have higher emotional intelligence than men.

[SPEAKER_00]: So are we looking at the age of women now?

[SPEAKER_02]: Ah, such a hard topic.

[SPEAKER_02]: I feel that we women are not, at the moment, not fighting for equality, but we are taking revenge.

[SPEAKER_02]: And I don't like this at the recent times.

[SPEAKER_02]: So...

[SPEAKER_02]: I'm not sure if women are really emotional, more intelligent than men.

[SPEAKER_02]: I think that men are very, so historically, very often stressed by their jobs while it was the jobs of the women to hold the family together.

[SPEAKER_02]: And that's why we have this impression

[SPEAKER_02]: And I think that there is some really some biological binding of a woman who just gave birth to a child, which cannot be that it's way more intense than the same with the father.

[SPEAKER_02]: So that's my impression.

[SPEAKER_00]: There's also a kind of evolutionary argument that I've heard been made when it comes to men being the survivors of people who have also survived different wars.

[SPEAKER_00]: And if you're very emotionally susceptible or sensitive, you probably won't survive a war.

[SPEAKER_00]: So then the hard men prevail.

[SPEAKER_00]: And that kind of evolutionary makes for less emotional men.

[SPEAKER_02]: On the other hand, I mean, women suffered so many loss of children in the former thousands of years that you also had to be very emotional, strong to cope with that.

[SPEAKER_00]: Yeah, that's a great point.

[SPEAKER_02]: I think that it's never a fight, men against women.

[SPEAKER_02]: Yeah, of course, a lot of jobs are socially occupied there.

[SPEAKER_02]: They're made by women, and those jobs are, at the moment, seems like those jobs are safer than the computer science jobs, yes.

[SPEAKER_00]: it might be a good thing for humankind I would think that emotional intelligence has a higher value yeah at least that's kind of a utopic future where that's like the most valued characteristic I think what's really interesting is that a few years ago if you as an artist said I was choosing this

[SPEAKER_02]: um object or this color or this uh melody because i really had the feeling that

[SPEAKER_02]: I love it.

[SPEAKER_02]: Or my emotion was the reason why I made this artistic decision that maybe the artistic world would look like, okay, this is a very female approach on art.

[SPEAKER_02]: I mean, art is an intellectual blah, blah, blah.

[SPEAKER_02]: But now it's funny that it becomes clearer and clearer that this is the main difference between us and machines.

[SPEAKER_02]: And this is the only important thing.

[SPEAKER_00]: Yeah.

[SPEAKER_00]: And what is the next step?

[SPEAKER_00]: I recently began reading about Zen Buddhism.

[SPEAKER_00]: And in that practice, it's important to free yourself from the mind and your thoughts and your feelings.

[SPEAKER_00]: And that practice is like the ultimate expression of a human is not connected to thoughts, intelligence or emotions.

[SPEAKER_00]: It's just a state of being, which is a spiritual thing that you can't describe, basically.

[SPEAKER_00]: Yeah.

[SPEAKER_00]: So maybe that's the next thing when the machines also can express emotions just as well as us.

[SPEAKER_00]: We'll go to the next thing, which is just the state of being in a Buddhist kind of sense.

[SPEAKER_02]: Yeah.

[SPEAKER_02]: Interesting concept.

[SPEAKER_00]: Something to think about.

[SPEAKER_00]: One last question for you, Esther.

[SPEAKER_00]: Do you have any advice for the future generation when it comes to reconciling authenticity and also being with the times and using new tools?

[SPEAKER_00]: How should they make an artistic expression or artistic career, do you think?

[SPEAKER_02]: This being authentic is so...

[SPEAKER_02]: so important and you have to be so brave to be authentic but it's worth it not only that maybe as an artist you will gain more followers but also that you can live a free life and this is worth everything nowadays yeah and

[SPEAKER_02]: Be honest and be kind and be brave enough to say, I use this tool in this way.

[SPEAKER_02]: Yeah, try everything out and be brave enough to say that something worked not out and now after three years you would do a song in a different way.

[SPEAKER_02]: Yeah, be brave.

[SPEAKER_00]: Yeah, be brave.

[SPEAKER_00]: Nice.

[SPEAKER_00]: Thank you so much for talking to me today.

[SPEAKER_02]: thank you so much for having me it's been a pleasure