Miss AI Podcast
I sit down with founders, builders, and fascinating people to talk honestly about how AI is shaping their world and what's coming next.
Miss AI Podcast
He Refused the World's Biggest Photo Prize Because He Made It With AI
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Boris Eldagsen has spent 25 years studying at the world's most prestigious art academies, photography in Mainz, conceptual art in Prague, fine art in India, philosophy at Cologne and Mainz.
In April 2023, he won the Sony World Photography Award with an AI-generated image. Then on stage in London, he refused the prize.
Today he's a Creative Partner of OpenAI, Runway, Higgsfield, Dreamina, and Hailuo. He advocates for the word "promptography." His message? AI doesn't suffer. AI can rant, but it can't bleed. The difference between human art and AI generation comes down to one thing, mortality.
In this episode, Boris breaks down:
– Why he refused the world's biggest photography prize (and why Sony's silence forced his hand)
– The word "promptography" and why AI images are not photography
– His full workflow for the Francis Bacon AI Art Rant (20 steps, 7 tools, 6 weeks)
– Why he uses AI to attack AI in his installation "This Is Not Art"
– The four levels of how art affects us — and why authenticity is overrated
– His honest take on whether AI is unfairly exploiting the work of human creators
If you're an artist, photographer, or creative wondering where you fit in the AI revolution, this episode is for you.
CHAPTERS
[00:00] The refusal that stopped the world
[04:00] Why AI is forcing us to define what's human
[07:00] How much of "The Electrician" was AI vs 25 years of being a photographer
[10:00] Where Boris draws the line between AI generation and art
[13:40] Promptography and why the world needed a new word
[14:45] Mortality, suffering, and what AI can never have
[17:30] The four stages of how art affects us (and why authenticity is a myth)
[23:50] Should AI-generated work be labelled?
[26:30] When will an AI movie win an Oscar?
[34:45] How do you know when an AI artwork is finished?
[37:20] Inside the workflow of "This Is Not Art"
[51:55] The contradiction, critiquing AI while being a Creative Partner of OpenAI
[55:30] Where to start if you want to make AI art
[01:02:45] The biggest lesson from 25 years as an artist
[01:09:50] What AI should never take from us
🔗 Follow Boris:
Instagram: /boriseldagsen
YouTube: /@boriseldagsen
Website: https://www.eldagsen.com/
🔗 Follow Keira:
/realmissai
/keira-nesdale-b287899b
https://www.aiconsultancy.co.nz/
🔗 Subscribe for more conversations like this:
/@realmissai
#ai #artificialintelligence #podcast #aiart #promptography #boriseldagsen #openai #aiethics #photography #generativeai #aiandcreativity #futureofart #digitalart #aicreators #contemporaryart
Graphic Refine is the old Greek word for uh writing and photos means light, so photography is writing with light, and promptography is writing with prompts. I have been working with AI for four years, I see a new development, I see a topic that most people don't talk about, and that becomes then the topic of my uh keynotes. One time I got the award from Sony competition, and I was totally surprised. I did not expect to get that far because especially with Sony, I have applied five times with photography and never got anywhere. I made a choice uh to refuse the award. My future bet with AI, anyone that has no artistic creative skills, background, education can get on a statistical average level with AI. So that bottom of creative production, yeah, of people that just don't know better is going to disappear.
SPEAKER_01Boris has spent 25 years studying at the world's most prestigious art academies. Photography in Mainz, conceptual art in Prague, fine art in India, and philosophy at Cologne and Main. He has built a career deeply rooted in the history of image. In April 2023, he changed the world of photography forever. He did something nobody had ever done. Boris submitted an art to the Sony World Photography Award, the world's biggest photography prize. And he won the creative category. The image was called Pseudomnesia, the electrician, an image of two women from the 1940s. But it wasn't a photograph. It was made with AI. The Guardian called your image the picture that stopped the world. The age called you the man that lifted the lid on Pandora's box. Welcome to the Miss AI podcast, Boris. Take me back to that moment. Why did you do that?
SPEAKER_02I saw a moment in time where I could make a difference. And I was one of the first in the German photo scene to play with AI tools. And very early in the 22, I saw the possible consequences. And uh having worked with photography for many, many years, uh, each autumn is the time for open calls, for photo awards, and I have been applying uh regularly. So I had a look at those photocompetitions. Did they change the guidelines? Because since September 22, it was up in the media that AI generators uh could uh produce images. And uh none of those photocompetitions changed the guidelines. So I thought, well, let's make a test. And I chose three of my favorite images and applied. And with the electrician, uh, three times it was shortlisted, and uh one time I got the award from the Sony uh competition, and uh I was totally surprised. Yeah, I did not expect to get that far because especially with Sony, I have applied five times with photography and never got anywhere. So uh then I thought, well, what do I do now? And the first two competitions where I was shortlisted, I did a post on social media linking um to the jury members, telling them that it's not a photo, and there was no response, there was dead silence. And uh with Sony, um, it was more or less the same. I told them that it was a test and that they could disqualify me, or if they would like to continue to have an open debate, and they didn't really respond. They said, Oh, it's fine, yeah. And um, I was very disappointed that they didn't make a choice. Any choice would have been okay, yeah, to say, well, we consider it photography or it's not photography and uh bear the consequences, but they just couldn't bother. And uh when they didn't uh make it transparent to the press that my image was not photography, um, I made a choice uh to refuse the award. And uh I was surprised that it became global news.
SPEAKER_01Yeah, global news everywhere. I even read an article here in New Zealand on New Zealand radio that was talking about you. So it was a definite show stopper. But I thought about this, and what it actually did is it brought into the media's attention the contrast between AI-generated real things and um real things. Like, what is the difference if judges at the world's most prestigious photography academy cannot make a difference between AI-generated image or an image that has the energy of a human into it? Where does this lead the whole entire world for deceiving people of what is real, what is not, fake news? I mean, think that it doesn't just stop at photography, it extends further into the other arts of writing books or text, also music generation. There's a massive debate going on in the world right now around what is AI generated versus what is not AI generated. How do you feel about that?
SPEAKER_02Well, yeah, totally true. Yeah, it's it's all of those big questions, and uh we're still trying to understand what is happening. Uh the positive thing for me having a background in philosophy is that AI is forcing us um to understand our human nature a little bit deeper than most people would uh dig into. Uh so all the terminology that is used, intelligence or creativity or consciousness, are topics for philosophy. But if you go into sciences, um there is not one definition. There are uh approaches, but uh we we can't agree what it actually is. And uh now everybody is afraid that AI is going to take uh these areas um over. And uh I have been uh working with AI for four years. I'm researching it, I follow the newsletters, I teach it. That helps me to dig deeper than I do a lot of keynotes, and every seven, eight months I see a new development. I see a topic that most people don't talk about, and that becomes then the topic of my uh keynotes. And um most people still haven't figured out the difference between art that was created before AI and today because uh it is complex and it's going to be more complex in the future. Um, on one hand, you have an artist who works without AI. On the other hand, now you have AI agents that are autonomous, yeah, they need a kickoff, but then they do whatever they think is right. And in between, you have all kind of uh uh labor division between artists and AI and different roles, and that is not just for one project. And uh how we will see later is that a project has many different steps. A workflow can easily have 10 different steps. So for each of those steps, you can decide again. Am I going to be the director? Do I have the creative lead? Or am I going to outsource it to an AI agent? And um, that is set for me. The collaboration between humans and AI is set also in creativity. And uh we need to figure out how to optimize this teamwork.
SPEAKER_01Talking of collaborating with AI, so the photo, the electrician, well, the AI-generated photo electrician that you created that won the award. How much of that would you say is your knowledge of being a photographer and studying art and philosophy for 25 years? And how much of that was AI?
SPEAKER_02I would say all of it.
SPEAKER_01Wow.
SPEAKER_02Well, um, I would say um all of it. It's 20 steps. And um the creative workflow in principle has three um different parts. The first part is the idea that I put in a prompt. And here I work for my professional background. The more you know, the better you can describe an idea, the better you can prompt. Then the actual production is outsourced to the AI. But then comes the evaluation. And here you need context, you need experience, you need to have taste, and you need to identify what is missing. How can I improve the prompt, or how can I improve the workflow? So from step three, you go back to step number one and you do many loops. And that image was created in over 20 different steps. And each time I made a decision built on my experience. And so um in the past, I have been working with AI in very uh clearly defined roles. I decide I'm the director, I'm the conductor, and the AI is my assistant, and it's uh many assistants uh and it's a tool. For me, it's a tool that I always wanted to have without knowing it would be possible.
SPEAKER_01You say that you're the director, but when you're putting a prompt into an image generator model, you don't actually know what's going to come out of it. You could be thinking a different sunset and it just creates a sunset. And when it creates the sunset, you say, Yeah, that's what was in my mind, or that wasn't what was in my mind. And then you also make the decision whether you like it or not. If you like it, you go with it. If you don't, you get angry with it and tell it to regenerate. So, how are you still the creative director? Because you're prompting it, but you're not always sure what you're gonna get out of it.
SPEAKER_02Well, at first I don't get angry. Even if you work with people and you're a director and you get angry, I don't think you're a good director. Um I'm the the the tools in the beginning in 22, um, it was hit and miss. It was like summoning some ghost, and you don't know who is going to arrive. And uh today um the technology has improved dramatically. Now you can really direct. And uh the the the prompt closeness um and um the the the way your prompt is uh translated into images is so accurate now. So that that has changed. And also, what is so different from how I worked before in photography? Um I staged images and I also uh went to locations always at night looking for opportunities, and then something happens, somebody just comes along, and then I see a kind of order in that chaos, and I need to click at the right moment, and then I have frozen this moment in a photography, and I also make a choice, and also stuff is going to happen that I could not foresee, which uh is the beauty of photography for me.
SPEAKER_01That's one of the arguments that you are talking about with AI-generated art versus photography or it or art that a human has created in with a paintbrush or with a camera. So, where do you draw the line between something that is AI generated and not, and considering it's still both art?
SPEAKER_02Uh are we talking only about photography or art in general?
SPEAKER_01I would say art in general. Um But we can know we can stick with photography wherever you like.
SPEAKER_02Well, let's let's start with photography and then and then then uh zoom out. Um photography, as the name defines, is an image uh created with light. Light is bouncing back from something in the world. And uh AI-generated images um are created with prompts. This is why I call them promptography. And um I don't need to be in the world or interact with the world. I need a computer and electricity, and that's it. So the strength of the AI is it's a very good tool for an inner journey, but also it forces me to be aware of who am I, what do I want to create and why. Where in photography you can basically just travel to a different place, you see something, and then you react, and you don't need to have any clue of who you are or what is happening around you. And uh the mistake that many people do, um, especially in photography, is considering AI to be a next evolutionary step, but it's not, it is just um one part of the big new thing, and the big new media is latent space. Latent space is um the training data, all different formats translated into vectors, and that is what we are working with, and with one and the same prompt, I can decide which kind of media format I would like to generate. I can generate the text, I can generate a sound, a video, an image, and and latent space is multimodal. And that would be the big claim for me that most people haven't realized that this is what it's about. They still come from music, image, video, whatever, text, and only use AI in that segment. But it's much larger. And that also means that uh promptography for me is much larger. It is also generated video, text, sound, etc.
SPEAKER_01The word promptography, you are a strong advocate of it. And is this a what does it actually mean, the word promptography? And why do we need that word in the world of AI and photography colliding?
SPEAKER_02Um well, uh graphy grafine uh is the old um Greek word for uh writing, and photos means light. So photography is writing with light, and promptography is writing with prompts. And uh I think it's very important to come up with a new terminology in order to differentiate, not to put it into one basket. There are several attempts to name the new thing in images, it's AI images, then uh there is syntography, synthetic images, promptography. None of them has taken the lead now and is generally adapted. Uh they are all running in parallel.
SPEAKER_01A classical test that you have described is that art is defined by intention, skill, and effect. And you mentioned that one thing that AI, like AI can do all of those things very well now. Um, AI can provoke emotions in photography or different artworks that it generates. It's getting very, very good. Like you said, the models are getting better and better every single day. The latest OpenAI image generator model just blew my mind. I had a great time playing around with it. But you said that one thing that AI cannot generate is that it lacks mortality. And this is where you brought the meaning back into the human actually creating the piece of artwork, right? Where if a person is creating something, they understand that life is not infinite. It has a finite destination. And do you think that that feeling creates different types of artwork that AI cannot understand or generate?
SPEAKER_02Yes. Um, we are locked in one body, one existence, personality. Uh we are born in different cultures with uh parents with many advantages and disadvantages, and that is shaping us. And we we can't escape that. And AI has no personality, it it can mimic all different characters. Yeah, you can subscribe to AI any role, um, but it doesn't matter. It's all of this and nothing. And that is uh creating a different need for humans to have an outlet, to communicate with uh other humans about what it means to be alive and and uh the positive and the negative sides. And uh art also helps us to understand what is going on, what I'm feeling, thinking, what I have experienced. And um it can also help us to um get different perspectives on life to to uh transform lives. And uh for AI, none of this is um important. Um recently um I worked on a new keynote that is uh looking much more closely in this. Um and be because in in the creative field, also in the creative economy, everybody talks about authenticity. Like authenticity is the anti dot to AI. And I had a closer look at it over two months and uh I came to the conclusion it's not. And if you're interested in that, I can I can take you through that. Um authenticity can be understood in many different ways about provenance. Is that a painting from Rembrandt or not? Yeah, then you try to date it back and trace all of this. Uh it can also uh be um an artistic um expression that you feel there is one single person and one life that had a need to write stories like Kafka, for example. It's very much bound to his uh biography. And then when you have this um sender, there's a certain relationship and trust you can build up. You can also understand uh authenticity as uh factual truth. Did something happen? But if you talk about literature movies, for example, um, we don't ask that question. If we read a story from Franz Kafka and somebody's waking up in the morning being an insect, we don't think is this real? Yeah? It doesn't matter. But um if we have a look on how like movies, art, advertising, entertainment is affecting us emotionally. I see four different stages. And the first stage always starts with a very strong stimulus. It's an emotional trigger, it's like a wow or an anger that is triggered. And we have been trained by social media that this is needed. If there is no emotional trigger in the beginning, people will not look at it. But that emotion that is the hook. Yeah, so we are hooked to the hook. And uh the the the pro well, on one side, that can fade away quickly. Next hook comes and the next hook comes, and I have forgotten what I have seen in my timeline uh 20 minutes ago. But if you want to go on a deeper level, it's about resonance. You are listening to a song and and the lyrics, and you think, well, this song is talking about my relationship. And then uh there is a resonance you you identify with um that situation or a certain community because it's an echo in yourself, and that is very, very important here. But this is something that I um just kind of like um put into it, it comes from my side. The third level is then uh a relationship that you're listening to the song, and then you think that person was writing that song for me, that person feels me. And you um you try to to um to research who is that person, what has that person done before, what uh uh yeah, you do you follow that person uh not to miss what uh uh is happening in the future. And here there's a very interesting divide between arts and entertain uh like fine arts and advertising entertainment. Advertising entertainment, also with AI, is aiming for frictionless whatever. Um with AI, especially in marketing, uh you will get different advertising than I. Yeah, it will work perfectly on you, it will work perfectly on me. Um so we get what already uh has an effect on us, but better. And that is what everybody is aiming for. Uh uh in the art world, it's different. Um, something works on us if I haven't ordered it, if it surprised me. It's somehow it it's Sticks in my head and it creates a conflict. And only if you have this friction, it can transform your life. Then that song that I have been listening to can change the way I think about my relationship. I behave in my relationship. And I think this is the top level of what art can do. And then if we come back to authenticity, the big question is: do we need authenticity for stimulus? No.
SPEAKER_01I agree.
SPEAKER_02Do we need it for resonance? No. I was at an event recently, and the host is a musician, and she was very angry about an EP she really, really loved, and she thought this was a young woman of color. And at the end it turned out to be AI done by an old white guy in Las Vegas. It shows that in order to resonate with you, doesn't matter if it's authentic or not. It becomes important if you build up a relationship and you want to dig deeper and you look where is this person coming from and so on. But even that today can perfectly be faked with AI. AI can imitate all the signals of the authenticity. And so if you would like to trigger those first emotional levels, it doesn't matter. Yeah, if if you are human or AI, if it's generated or not. Only if you're in the art world and and um you would like to to come up with a stronger relationship, then it becomes important. So what is more important than authenticity here is transparency. How is something uh created and friction? And then where does friction come from? Yeah, who is creating something that is not streamlined? And then it needs to be a single human being, yeah? It needs to be something that comes out of your daily life, your experiences, uh, and then authenticity comes back. But uh it's not as important as most people think. And I think in the future, friction um will be something we need to work for.
SPEAKER_01Do you think that all AI-generated work should have a mandate that at the bottom of the work it should be declared that this is artificially generated?
SPEAKER_02No. Um I don't see any option how you could do that. I think um people should be just insparing about your workflow. Um it's very important that um in the news, in the photojournalism and journalism itself, we have a certain level of authenticity. That would be easier to create, and everything else doesn't matter. I've been um part of a working group of the UNESCO World Heritage. There's an immaterial world heritage, and since a year now, we debate the future. Is it possible to say something that was AI generated becomes UNESCO World Heritage? And uh the question is then authorship. And there are some people that come from media theory and archives, they want to preserve hardware, software. But in AI, the companies will not give you their algorithm. And uh artists will not write down all the prompts they have been using. So um what I see right now is we are in this in-between time where this seems to be necessary for us. It seems to be a solution. But if you look um into the future and if you realize that AI is an enabling technology like electricity was and is, when electricity was invented, we had no idea that it's going to power the way we connect now, yeah, from one side of the globe to the other. And we will have no idea what AI is going to lead us to. We don't have a category for things that are created with electricity. So in the future, it doesn't make sense to have a category for stuff that is created with AI. And um yeah, that's I I was growing into that thought. Yeah. In the beginning, I also thought we need to draw a line. When is the input of the person enough to say you are the author? And when are you just typing two words and that's it? But practically, I don't think this we can manage to draw a line there. It just doesn't work.
SPEAKER_01Yeah, I've thought about this a little bit when the image generation, actually the more so the movie generation, get started to get really entertaining and really good uh over the past couple of, well, actually on the past six months, it's becoming really, really um captivating. And it made me think, I wonder when a artificially generated movie, a feature film three hours long is going to win an Oscar. And then I had to think a little bit deeper around like, would it still be in the same categories as traditionally filmed movies? And I couldn't answer that myself. I was sort of frozen in headlights and I didn't really know. And I don't think we know yet because we'll get the first one, and that'll be the disruptor of the industry when it when we do get that three-hour-long feature film that is Oscar-worthy, and actually giving the traditional ones a run for their money. It's not there yet, but I think it's coming.
SPEAKER_02Oh, for sure. And there are already many uh feature films made with AI, TV series, uh short films. And I've been working with AI video for two and a half years, and I I truly love it. Um, I've done videos since my studies in the 90s, and the latest project I did was called Super High. It was a mockumentary casting show on getting high with no drugs and a jury on drugs judging it. And that was actually co-financed by the French-German uh cultural TV station Arte. So it was state money. And um, but that took two and a half years of my life to do everything, organize it, yeah, um, direct, shoot, audio, editing, and after that I said no more. It's I go back to photography, it's more instant gratification. But now I'm back, yeah. I do more videos these days than than I do um still images because it's much more efficient. And uh the latest video I've been doing, um they are short, they're two to five minutes, but um uh I have done videos in two days, yeah, or five weeks, and all of this is so much more effective than uh going for two and a half years. So um that is something I love truly about it. Um, and many uh film festivals, and I've been applying to film festivals with my work, um, they have opened up a new category. So most of them have an AI category now. I think this is where uh the Oscars also will be heading. But uh there's one thing that is really interesting. What is the big difference? And um, there still is a certain magic of acting that comes on set that is translated onto screen. That is that if you have professional actors interacting, they can take you into any kind of emotion and they can express any kind of emotion that is a skill. So are we only watching movies to be entertained or to be touched, or are we also admiring certain actors and the skill they have? Um I think AI will be able to copy this as well. You already have the tools, yeah. You you can use any actor, you you can record you and you and I, and then with AI we become different characters. Yeah, we can we can be, I don't know, Lord of the Rings, part 20, and I'm an orc and you're an elf. And and um that is easy. But if you then have someone um that can really act and express um to turn into those characters, um that is a different level.
SPEAKER_01So what also say, what makes it award-winning? Is it the storytelling? Is it like still what wins these those awards? Because you've been in the industry for over 25 years. Can you tell the difference when there is a artificially generated video or I mean the videos are a little bit different to to tell, but maybe not anymore. Hey, how do you tell?
SPEAKER_02Well, I I I've been in juries now for AI images, not that often for AI video, but um, my initial response to what makes something award-winning would be the jury.
SPEAKER_01And that's just like traditional media, then you can buy it.
SPEAKER_02It means it it depends on personal taste. Sometimes it depends on connections, which is wrong, but this is how it works. And uh I have been in a jury uh at uh photo festivals that opened up AI categories with other jury members, and they had no clue. They had no clue how AI images were generated, and they uh just said, I'm going to judge like in the last 20 years, and I can say, How can you? You need to be aware of the process to know that this image that looks like a Magritte painting as a photography, it's a landscape, and uh, then there is a chair, and there a person sits in the chair, you see the back of the head, you see the elbows, and uh then this is a screen in the landscape. So that was like two years ago, and uh I said it's it's an interesting image, it's a reference to art history, to to surreal painting, but you don't see the face, you don't see the hands, and this is actually difficult. That was difficult that time to generate. And here somebody is just kind of like having um sneaking out, don't want to confront the problem and and uh choosing an easy solution, not showing hands and faces. And that is something that you need to keep in the back of your mind when you think this is the most brilliant image. Uh so it always comes back to the jury, their knowledge, uh, which is a kind of gamble. And on the other hand, it also depends on uh certain trends and also where this jury is coming from. I've been in the jury of the AI art magazine now three times. Uh, and uh in in the in the first two times I've been the only artist, and most of the others came from advertising design, and they made different choices.
SPEAKER_01Different perspectives, all right?
SPEAKER_02Yeah, they look more like something like uh visual styles, and for me, that's just like half of it. And for me, art is somewhere between beauty and the disturbing. And for the last edition, uh the jury members could give uh a golden ticket, and that means a person is in the magazine, doesn't matter how many votes. And so I gave that golden ticket to an Australian artist, Ian Haig, who is specialized for over 20 years in body horror. But that means what can go wrong with the body? Because everybody is showing like like pretty sexy cyborgs, yeah. But what if all of this goes terribly wrong? And and he's working on it and it's very strong. And I got him in the magazine, and most of the others would just say, How dare you? This is so ugly to look at. But then the question is, why does everything need to be beautiful? Yeah, why don't we want to confront that our bodies are aging and decaying, and then this is part of life? And I also see his work as a psychological portrait. Yes, uh such mutated bodies will not survive, but uh many of us look like this on the inside from time to time. So um yeah, now we got into different uh territory.
SPEAKER_01No, that's okay, but so if you're an AI generated art generation artist, how do you know when your AI generated work is finished?
SPEAKER_02It's uh like before. You just know. It's it's it's a kind of gut feeling. Um before, especially with film, you also had material restrictions. Like, okay, this is the material, I have no money to shoot it again. You need to work around it. Today you can reshoot. That is great. Yeah, you can just do replace scenes and and change scenes, but again, yeah, at some point you just feel you look at it and it's right. Yeah, the flow is right. The feeling.
SPEAKER_01You just know.
SPEAKER_02That's the I don't know. How can you teach something like this? It's uh just by doing the yeah.
SPEAKER_01It's the artist's perspective comes into it, right?
SPEAKER_02When a little more dull or like you just know, and I think you can see that when um yeah, at the end, uh the it's it's the the difference between being an artist and being a designer. Yeah, or uh if if you work in the creative economy, you have a a target group, you have a task, there's something you you need to beautify or to sell. Uh as an artist, um I don't have a target group. I basically start doing stuff for myself. I am my first audience. And um, that was very uh always very necessary and important to me because uh when you decide to become an artist, it's it's a hard way. Yeah, you have need to have multiple jobs, and you look, you see all your friends in the same age, yeah, having cars, having houses, having families. And uh if you really want to be like a 100% artist, it's more like becoming a monk and going to a Buddhist monastery or something like this. And um then the only freedom I have is only to create what I think is right. Yeah, what I think is important. And even if I don't sell, it doesn't matter. Yeah. For me, for myself, I have succeeded. Success here doesn't lie in money. But of course, you can also see being an artist as a job, and then you follow trends and you create products for any trend that is just in fashion.
SPEAKER_01That is a great segue into your art installation. This is not art. So you have just said that you create art for yourself, and this installation was inspired by dead artists who are returning in a punk theme to attack AI. And everything within this exhibition was created by AI from the voice, the music, the imagery, the videos, everything. So, would you be able to take us through your workflow and creating this, these artworks? Um, and yeah, just just how do you go through that whole entire creative process? I think there'll be so many people that would love to dive deeper into your mind for that workflow.
SPEAKER_02Yeah, it will be a pleasure. So, uh, as I said before, uh I'm doing multiple things. And uh after uh having spoken about the difference between photography, AI images, and the effect on creativity, I started a series called Um AI Art Rans, and this is not art. It's an installation where that artists and art movements complain about AI. Yeah, so you can possibly put a link into the show notes. And um that is how it looks at the moment uh in Norway in the museum. It's uh 10 meters high, so you have uh five videos, and then you have uh famous artists of the past life-size cutouts on the wall with one quote. What is the difference between their art and what AI can simulate? So I take you to the workflow of the latest video, which is uh about Francis Bacon, uh British painter who was very influential from the 60s to the 80s. And um he painted um screams, he painted uh human suffering, he also painted animals and slaughter horses and uh houses and and faces that look, I don't know, it's hard, but it's very unique. And so how how did I do that? So uh uh he was one of my favorites when I was not a student, so I I get got back to him and researched uh how did he work? So that is how his studio looked like. Yeah, it's a total mess. Uh all of this was in London, now it's in Ireland, in Dublin, it was turned into a museum, you can visit it. And then I collected images of him and all of this as reference material, how did he look like, how did he work? And uh I watched documentaries. Uh, there is one from the 80s, uh, from a British um documentary maker, which is great. You see uh one scene in a restaurant where he drinks, I don't know, one and a half bottles of red wine. He was also drinking a lot of champagne. So all of that got me into the mood of the person that is his paintings, he's favorite for. And then I started with the monologue. This is in the beginning. And that was with ChatGPT. Um I also use it to research certain topics and what uh quotes. And then I said, okay, I focus on a monologue about screaming and why AI can't scream, and what is the difference to um us humans? So what you do here is you create many of those monologues, like 40.
SPEAKER_01And You're having a conversation with OpenAI or ChatGPT and uh going backwards and forwards and asking it well questions about your research?
SPEAKER_02Um well, first the research, and then it's the monologue. It's it's a monologue where I want um uh ChatGPT to imitate um Bacon's style of talking, and uh I define um the topics of the monologue. And then you get different versions. None of them you could actually use, but in each of those versions there is one brilliant sentence or half of it. So what you do is edit, yeah. You you cut and paste, and then you put all of this together. And uh I I do this always when I'm traveling. So on the left on the black screen, this is this is my note app from my smartphone. And I was doing this when I had a 14-hour flight to Shanghai. So um and uh once I have the feeling, okay, this is ready to go, uh I create a kind of like um storyboard and an order. Um how does it start? How does it end? And uh you see there are certain chapters, and then uh from those chapters, I just create images. I have no clue. I just say, okay, he's painting, he's smoking, he's drinking, he's screaming, he's destroying a canvas. So I create and create. And here I always compare different platforms, and the leading platform is always changing. So at that time it was uh Nano Banana Pro and a little bit of C-Dream uh 4.5 that I used. So I had those images.
SPEAKER_01Can I have a question? Do you have a story arc in mind that you're going to create? So if you're thinking of creating a three to five minute video, do you know in your mind where that's going? Or you kind of just like you know, piece it together as you get little snippets from the monologue, and then you know, you have researched around this artist, you know that he likes to drink champagne, and then you sort of like how do you create that story arc?
SPEAKER_02Well, a story arc is predominantly the monologue for all of these series. What is changing now and what has been changing with uh Francis Bacon tiptoeing around the void is that the visuals become more important, and then uh uh something else enters the scene. Like the the first two videos, Bukowski and the Punks, which I still love. Very easy. It's just punks being angry. And I I love the monologue and everything they say about it. It's called Tech Pro Vomit in HD. It's just that. It's very easy. Yeah. It's screaming into your face, and the monologue is the story and then finished. Here it took me much, much longer because I wanted the visuals to have a life of themselves. This is why it became twice as long. But it is driven by the monologue. But this is only one way of working. Right now I create video for a concert series of friends. They are contemporary composers, musicians in spatial audio. And I'm the only visual artist, and I don't even know what they do. The second concert is on Sunday. And last time they created sounds with anything electronic, analog. Beautiful to watch. So what I do there is because I don't understand what music they are going to play, the topic of the night, like next time it's it's time, temporal compression, then I improvise on it. And then there's no text. I basically create images, and then with the text prompt, I make the protagonists in the video do an audio performance. Working with C dance is very effective, working with sound. And then I have videos where there is no spoken word. But it's a kind of like imagined audio performance of people and objects. It's a different way of working. But here's the monologue. And um after I have created all of those images, I try to number the monologue. I try to number the images, I put them in folders. And then the next step is uh animating. And then here's one example. I I hope it it it uh comes. Yeah. I don't know if you heard the sound, but it's a good one.
SPEAKER_01I didn't hear the sound, but I could definitely see the artist screaming. Uh it's good you didn't hear it.
SPEAKER_02It's it's very loud. So that that was one of the outtakes I didn't use. And then you can say, okay, do you want to have a camera movement from him to the painting that he's famous for? And you try out this stuff, and then the the still images are little videos. Um you edit conventionally. I'm using uh Premiere Pro of Adobe, and then uh you see you have different levels of images and sound, and the sound is always 50% for me of a video work. And in the process, you realize um I need more material. There is something that is missing, or you're rearranging the scenes and you're changing the monologue or cutting off parts that you don't think are needed. And then you generate again material that has been missing. And uh here I give you an example of how this process is something that you can't foresee. Um I know he was a heavy smoker when he painted, and his studio was a mess. So I think, of course, he's setting it on fire accidentally. And then you think, okay, now I need to extinguish the fire. What is he doing? So I tried out pouring wine over it, champagne, spitting, pissing, and then the pissing became the best scene. And um, what the AI did uh after that that that fire was extinguished, suddenly I had this hole in the floor, which I didn't prompt. And then I said, wow, that's great. So I have the pigs that are appearing on and off in the video. So let it appear here. So there's a space below.
SPEAKER_01I didn't even see the peg and tone right now. I just saw it. I would be looking at that hole for the last minute. I didn't even see it.
SPEAKER_02Yeah. Um, and and these ideas, um I I can't know before. I can't outsource to AI. It's the same working without AI in photography and video. You produce ideas, you refine ideas in the process. And this is why it's really important for me to use AI as a tool, because this is my handwriting here that enters it. And also Fox uh appeared in in different forms. And then the problem that you have until now still is the voice. Um there you can uh have uh consistency with the faces using reference images, but uh the best models don't have a consistency for voices yet. Um that means all of those Francis Bacons, they have a different voice. And of course, in in the punk video I kept it in and videos before, but here I wanted to have one voice only and also creating different sounds for the soundscape. So um I was using an old platform that exists over 25 years called freesound.com. Uh, when you you need the sound of a fox or of a pig, you find it there and you download it and put it in the edit. And uh I wanted to have a song because he loved listening to classic music. And this I invented a classical song uh singing scream, scream, scream. All I want to do is scream. Um, that is kind of the background, also AI generated, and then finalizing the edit. Um, you see how complicated it has become. And the problem of the voice is sorted in the end. That means you need an additional audio software, and 11 Labs is one of the best where you can generate your own voices, or you get one that is like uh preset, and then all the different voices I had in the videos become one. And for me, it's very important that it's not the original voice. I could have copied it, but it's different because I don't want to generate uh something that people think, okay, this is really him. There needs to be an element that is obvious. It's not his voice, it's a British voice, but not him. And then you export the edits, and then uh videos are generated usually in 1280 uh pixels, but now the first models come out with 4K natively, but you upscale it most of the time because generating in um smaller um uh radio is uh saving money and time. Then you can go to 4K, you could also go up to 8k, and then basically it's done. And then the rest is the installation um where you have similar um problems, and then you just arrange it somewhere. I usually have photos and measurements of the space.
SPEAKER_01Wow.
SPEAKER_02So that's it.
SPEAKER_01You just explained through at least, I think, seven different tools that you were using to create the whole entire video. How long did it does it take you to generate one of those videos? You mentioned this one took a little bit longer than normal.
SPEAKER_02Um that took longer because at some point I was stuck, and then in between I had the exhibitions and different stuff to to take care of. But um it was kind of boiling in the back of my mind since autumn last year. So that that might have been seven months. But if I condense it to how long did I actually work on it, possibly six weeks. Well, or less.
SPEAKER_01And also I want to address something. So watching you go through that workflow process, I kind of felt that I was almost pulled in two directions because you're creating an art exhibition that is 100% AI generated, but it's still art. And that's kind of a bit of a debate that we've been talking about for the last 50 minutes. And I also did a little bit of background research on you in that you're a creative partner with OpenAI, Runway, Higgsfield. You've been teaching AI art generation to directors and creatives and everybody teachers around the world. It kind of seems to contradict each other that you are in with the big leagues, working alongside them with their model generations, and then also on the outside, you are saying, like, this is you guys shouldn't do this.
SPEAKER_02Um I don't think it's that complicated. Uh I use um AI as a tool. And if you go back in the history of art, artists always adopted early to new technology. And uh because I have the lead here, um I I could have created these things also as paintings or as videos with actors. Um, but for painting, I don't have the skill, and for video I don't have the budget. And here the twist is that it's created with AI criticizing AI. I think what is really important is um to help people to understand the differences. It's not saying AI is bad. I think as an artist, I truly love AI. But I also truly love painting, photography, film, everything else. Um for me, it's important to know the strengths and weaknesses of each medium. And then it depends on what are you using it for. And I truly sympathize with all of the arguments of those dead artists against AI. I think it's right. Yeah, if the punks just complain about um video just uh looks generic like a wet Marvel dream. It is like this. This is what I see as a creative partner of all of those companies. In the channels, I meet the other creators. Yeah, it's usually Slack or Discord. What they want to produce is a Hollywood copy. Yeah, they they get off being able to do a fight scene martial arts where which is perfect, then they are happy, but there is no meaning in it. And um so I think those those arguments um I do support. And um when I'm teaching the tools, uh I often say to the people, um you need to incorporate it into your artistic creative life. You need to use your own skills. Only then you create something that is not a statistical average. Only when you are the director, you can go above. If you leave it all to the AI, what you get is mainstream.
SPEAKER_01I've done a little bit of AI video generation and also image generation, and I do really enjoy it, stitching together the little story, and I kind of just make it up on the spot. I made a video where it was myself. I took a picture of myself actually, and then I turned myself into an alien, and then I went on a journey through my house, and I didn't know where that journey was gonna go. And it was just sort of on the spot. I was like, oh, what would I do next? Where would I go next? Am I going into the bathroom? Am I going into the kitchen? Am I eating? Am I drinking? Am I smoking? Or like, what am I doing? Um, and it was really captivating. Uh, I spent way too long. I always spend way too long on the video generations because it's quite fun. Uh, it's very exciting. But and I'm also seeing a lot of uh video artist generators on uh TikTok and Instagram and they're on YouTube as well, creating awesome pages. And these videos are getting hundreds of millions of views. And so my question would go back to you is that if there are anybody who is listening out there who perhaps wants to get started in the AI image generation or video generation, like how would you guide them on uh yeah, getting started?
SPEAKER_02Well, the the knowledge is all out there. I think if you do a proper research, you can find how-to videos on uh YouTube. The main problem is to stay updated because the development is so fast. Yeah, often you have like the best tool this week, and the next week it's a new release. Um, one channel that that I like and that I follow regularly is uh Tim on YouTube. Theoretically, Media is his name, and he gives an overview about the new tools and what you can do with them. Um in the film world, you also have Curious Refuge. They do the same. They also do workshops. I've been doing workshops since 22. So I think the best option would possibly see what you can find for free online. And if you don't get enough information, look out for a proper workshop that gives you the basics, and then you you start from there.
SPEAKER_01Do you see one category of artwork uh standing out more commercially viable than other pieces of artwork at the moment?
SPEAKER_02You mean in AI?
SPEAKER_01Yeah, for AI. So if I wanted to get started doing creative AI artwork, would you guide me in a certain uh category of artwork, or are they all equal? If I was looking to make money out of that.
SPEAKER_02Well, if if if if you um want to make money out of it, then you're creating a product, then you're working in a certain creative industry. We can talk about that. Um if uh you want to express yourself and want to create uh art, don't expect to create money out of that quickly. Um that is one of the most difficult paths. It takes stamina over years, it takes a lot of networking. Sometimes it takes luck to be at the right place at the right time. But um whatever you do, I think you only enjoy being creative and producing stuff if you like what you do. So the first step would be to look into yourself. Yeah. What would you like to become? What would you like to spend your time with? And what I often did with the students uh when I was teaching in Australia was a kind of like um self-examination. Look back into your teenage years, what art did inspire you? Music, film, literature, etc. Write it down. What did you start creatively, possibly stopped, and why? Yeah, many of us played an instrument and stopped or did writing and stopped. And then do a map and see the connections. And um you will see um that you can put it in certain groups. And this is what you as a person are interested in unconsciously. And the beauty today is this basically pattern recognition. Now you can do the same process with AI. And I've done that last year when I created a show with uh an older comic artist from Sweden. He said, I can't write an artist statement. And I said, Well, let's have a Zoom. Tell me uh the five artists uh that inspired you, and then we put their names into AI and asked for pattern recognition. What do they have in common? And then I gave it back to him and I said, uh, which of those um uh topics do you identify with? And so we started to write his artist statement. I think this is very important for yourself. If you just like to be creative, but that's it, and and it's about selling stuff, then it's a creative economy, and then there are different ways of studying. But um, I know in in talking about this before, you ask for a first 1k playbook. Uh in in uh contemporary arts, no way. It doesn't work in seven days, and uh if if you are lucky, it works in seven years. Um if you go to a creative economy, um of course, you you can give some advice. So um I talk about photography because this is um where I know the most of it's like always you need to sell a problem, you can solve. And uh the last workshops I did with photographers, I said, What are your problems? We can talk about it. And AI is now very good in post-production. And then there are photographs that said, I photographed an architecture and there's a playground before, but it was a very hot day. There are no children. How can I get them into the image? Or I photographed uh an in a fair and there are people around the booth, but they are not enough. Can I put more in it? Or I photographed a person, and I realized later the shirt had weird creases, and can I get rid of them? So all of these things you can do. There would be a problem you can solve, and then you can create like three specific offers on the second day, and then I would not uh say, hey, I can create AI images for you, but I can uh create uh I can solve a problem in post-production that you might have, that a certain group of photographers might have. And then on the next three and four days, you create uh examples for this. Um, not that many, maximum 10, but strong. And then you set up a sales page and you show before and after and what you can do and how fast and a certain price, and then you need to find your target group and uh contact 50, 60 people on day six and explain what you can do for them. And if you are lucky on day seven, two of them are going to say yes if you have a startup package that is low enough.
SPEAKER_01I love that. Thank you very much for answering that question. And uh I want to know for you personally, you've been on an incredible artistic journey for the last 25 years, exhibitions all around the world, studying all around the world. Like, what has been your biggest takeaway lesson? Because you have seen such a transformation in the world of art going from well, just now that we're able to use AI to compete with traditional artwork. Like, what have you learned from all of this? And where do you think it's heading in the next, let's just say one year, because it's too far to predict five years ahead in this industry. It moves way too fast.
SPEAKER_02Um technically, I always said it's I can't predict. What I can predict is it's going to be better. And uh that's good. And possibly faster. When when people ask me what's happening in five years, I was laughing. I said, I said I can't predict the next five days. And uh, I was surprised in the last four years, and this is going to continue. Um, my future bet, um, the two things. Um let's talk about the education and and art academies, I think. Um with AI, anyone that has no artistic creative skills, background, education, um can get on a statistical average level with AI. So that that bottom of creative production, yeah, of people that just don't know better is going to disappear. And my bet is that in the future art students are going to go down there because this is a location they can squat. And then it's going to be something like a like a punk revival, yeah, where we see stuff that looks ugly and bad and on purpose. This is one thing. And the second thing, something that I'm involved very much, is how does AI affect um education, um universities, art academies, and I'm teaching since a year now in in Barcelona at the Labazart School of Art and Design, AI for creatives. Um they did the right move a year ago. They set up a completely new study focusing on AI only, where they teach to generate images, uh, audio, video, and uh this is targeted at professionals. So the people that uh study there, it's it's a year, it's two two evenings a week. Um they're between thirty and fifty. Yeah, they come from art, from design, from from media, and they are global. Student group, I had someone from Australia in there.
SPEAKER_01Nice.
SPEAKER_02And I think that is good because what I see most of the other universities they try to squeeze in AI in an existing curriculum. And that is not working. Yeah. And I'm now in the in the final selection of a new professorship in Germany on AI and multimedia in the arts. This is the first one that was established in Germany. In other countries, it goes faster. Recently, there was a center established in Norway researching AI and creativity. And that is a network that is truly impressive because they pump out 200 PhDs on the topic. And I think this is where we need to go. And if you look at other suggestions on how universities can change, in the past we studied a topic. It's like we go to an island and then we try to make ourselves comfortable with it and establish ourselves and build a house and a castle and so on. But today this pure um knowledge of facts that is outsourced to AI. And working with AI, we are navigating the open sea. So it's less about settle down on an island. It's more how can we navigate between the islands and discover new islands and continents and what are the skills for that? So one prediction would be that the way universities are structured is going to dissolve. It's less subject-centered, it's more skill-centered. And I think it's a very good question: what skills do we need in the future? If the production of creative stuff is outsourced and it's more about the idea and the evaluation and the context and the taste, how are you teaching these things? And as I said before, um the way generative AI works is creating stuff that is a statistical average. And so you need to teach people to go beyond. And that is basically system hacking. That needs to be a future skill. And that is my bet that it will take a long time for the university system to adopt, but they need to, and there will be some of them that are faster. And uh yeah, this is this is where I also see my future to be involved in some of those institutions that are faster and first.
SPEAKER_01Yeah, I agree that humans are going to become so reliant on AI. So you're right in saying that you're gonna have to teach the students to think, to be creative, um, on their own without any digital influences, which is a little bit reverse from I feel the trajectory is going. Everything is becoming digitized. Everything, even myself, I know that I'm always checking my phone or in front of the computer. And I actually even laughed to myself the other day. I was like, when was the last time that I really had proper original thoughts? Like, no distractions, just stared at a blank wall. And uh, it's an exercise that I actually have to do is like remove all my distractions because it is just so much going on in everyone's world today. And it's so easy to uh be connected with other people and not really follow what you're meant to be doing. So I agree. I think we need to, yeah, remind people to think and teach people to discover ideas without external influences, go back inside the mind, right? Which brings me to my last. No, I was gonna say it's probably a very fitting uh question to round out everything because we've been talking so much today around that AI can do pretty much everything for us. And it comes down to who is the conductor of the AI, who is the person talking to that AI and prompting the AI to create and generate. So a question that I ask all of the guests on here is that AI can do a lot of things for us now. What is one thing that you hope that it never takes from humans?
SPEAKER_02Well, my first gut uh response would have been consciousness, but then I thought it's not going to happen. Um, many people think that AI has consciousness, but uh we are not able to understand how it works in ourselves. And then we try to translate this into mathematics and algorithms. How should that work? Uh I'm very positive this is not going to happen. Um and um on a on a human level, um I also think feelings will uh not going to be replaced. AI is going to be incorporated where um robotics and AI are going to match. This is the next step in the development development that uh you have a a robot trying to explore the world to understand what gravity means. Um but from an artistic point of view, I think um I hope that AI never takes from us suffering. It's not that I that I I uh would like people to suffer, but I think um it is how we grow and how we we we also become stronger and and and work through it, and it also gives us a certain deepness.
SPEAKER_01And AI can't have that deepness, it doesn't have the experience to uh get that deepness, right?
SPEAKER_02No, it's it's always just simulating. It has the data about it, it has the effect and it it can echo it. But the the the really important difference is that all that we assume AI has because of its actions is just responding to our data and mimicking it because it behaves like us, it's not like us, and and as humans we always have a tendency to uh uh make stuff alive. Like how many people give their car a name or their computer?
SPEAKER_01They're giving their AI agents names now as well. I must admit, I've also named my AI agents, so I can't say anything. I've got my one called Theo.
SPEAKER_00Yeah.
SPEAKER_01But anyway, uh I want to say thank you so much. Today has been a really captivating conversation. I've been right in there, and the debate is very present at the moment, and deep fakes and the convergence of AI and art and photography is definitely a subject that people are starting to explore. And I think it's gonna become more and more. I don't know whether it's gonna become more and more debated because I don't think that most people actually realize the difference between AI generated art sometimes and not. Just for example, like I was watching a video with my my dad the other day, and he had no idea that he was watching an AI-generated movie. And I did because I could hear the tonality and the voice. But I don't know whether people will it will people debate this in the future, or is it just gonna stick with the creatives and the people that are traditional artists fighting against the AI-generated world?
SPEAKER_02Um I think some people will will not care about the differences. Um this is where being based in Europe, the European uh uh AI Act comes into play that uh forces um uh producers to make it transparent, is it AI or not? And that should be put into practice in August, if it's not postponed again. But that is something that um I think is also important. Uh we we talked about the different ways of how stuff is triggering us emotionally. Uh and I said uh it doesn't matter if it's AI or not, which is true. But there are many studies that show that if you know first it is AI and then look at it, um, most people have an anti-AI bias. Then um the level it's going to affect you is less. And therefore, I I think that uh transparency and that labeling has a positive side. So we'll see. Yeah, we should talk next year when it's put into practice.
SPEAKER_01Yeah, we'll definitely have more conversations. So I'll drop your sh your links down in the description. But uh verbally, how can people follow along on your journey?
SPEAKER_02Well, the easiest is Instagram. This is where I'm most active. Uh all the the talks and then video artworks I do uh I have on YouTube. Um I'm also uh on LinkedIn and X and uh Facebook, but most active Instagram.
SPEAKER_01And that's just your first name and last name?
SPEAKER_02Yes, very easy to find.
SPEAKER_01Cool. Well thank you so much. It's been an enjoyable conversation today.
SPEAKER_02Likewise. Thanks for inviting me.
SPEAKER_01Always.