Diolch yn fawr iawn am wylio'r fideo. Thank you. Thank you. Thank you. Thank you. Thank you. And I'm here. Come down. Thank you. Thank you. Thank you. Thank you. Thank you. So, welcome back to our last panel today, which is our beloved artist position where we present specially talented artists and significant artworks and we have the big honor to present today's three positions starting with Alex Werherst Alex Werherst is for us kind of known. We are happy that you are back again We had here at the pre-forum 2015 and you won the Golden Neeker and it was a really great talk and today Alex will present her new work Art Hominem, which is based on a theoretical work but I'll let it to you to explain directly what it is all about and what the experience we all had now in the last two years has to do with it. Thanks a lot for your coming. Welcome. So thanks for coming everyone. In March of 2020, the world kind of came to a screeching halt, as we all know. We were forced to pause and look around us, reconsider where were we exactly heading? What were we looking for? Is there a planet B and what does it look like? And will we actually be able to get there in time? I'm Alex and today I'll be talking about my philosophical video game at Hominem. So I am a very lucky guest professor at Lucas School of Arts in Brussels. I love being around 20 year old often brilliant minds that are trying to figure out the world. I'm always very happy to see the projects of my students which they make within the framework of my class, time-based art. In 2020, after a very challenging semester of online teaching, I was presented with projects that were either about visions for a better future or actually mainly about the apocalypse. One of the best students of the 2020 class, Ernest Tiesmeyer, made this quite epic video essay, a comparison between 15th century depictions of the end of the world and current actual news footage. Funny, and also quite frightening at the same time, as you can see. So this kind of made me wonder, where do these ideas about utopia and dystopia voiced by the 20-year-old geniuses come from? How do we think about a good place? And how do we imagine dystopia? I mean, we all had ideas about how a pandemic would look, right? I mean, it would involve a lot of running, alarm sounding, total societal collapse, maybe a zombie or two, right? But I guess before 2020 nobody could have imagined that the apocalypse looked like this. Or this. Or even this. So I started digging a bit within the pile of books I had lying around, now that I actually had the time to read them, if I wouldn't be anxiously scrolling through social media all day, but whatever. And I found Utopia Unbound, a brilliant PhD written by my good, dear friend Sophie Vrast. written by my good dear friend Sophie Vrast. In her doctoral thesis, Sophie defined four different style figures of utopia that circulate in Western literature. She investigates how the concept of a good place can activate stereotypical patterns of imagination, which in their turn circulate in a culture as a collective simplified framework used to understand a more complex reality. These stereotypes are the pair of glasses through which we see our current times, through which we understand our history, and by which we form ambitions around what we want for the future. I was very intrigued. I mean, there is a beautiful and frightening message to this research. It holds the promise of some common ground for humanity. Whether that promise is fulfilled, of course, is another question, a question I will try to address later. Per Utopia, she discusses three books, not necessarily the most canonized or best pieces of literature, but that is not the point, of course. With each selection, she shows, in fact, the spectrum of representation of these four ideas and the clear link between their respective representations. Her examples explicitly range from dystopic to utopic views on these utopia. So let me list the four style figures and the 12 corresponding books she quotes in her thesis. The Oasis and The Comeditarian Dream is the first one. Here she uses Aldous Huxley's Island, Chemin de Fer of Benoît de Teutre, and the new city of Stéphane Amidon as case studies to exemplify these utopic tropes. The capsule and the isolationist dream is exemplified by Melinda and Dragoman of Gheorghe Conrad, which is an immensely beautiful book. If you haven't read it, you really should. Les Quatre Murs of Ragon and The Fence of Van Looij. Then she speaks about the hub and the dream of futurism, which is quite pertinent to Ars Electronica, I always find, in which she uses the classroom of Marr, the architects of Hay and Super classroom of Marr, the architects of Hay and Supergun of Ballard, a very nice thriller, by the way. If anybody likes crime novels, it's a really good read. And last, The Bazaar and the Dream of Hedonism, exemplified by Le Parc, Arcadia, and another Ballard thrill, Cocaine Nights. So I loved the dissertation and read most of the books with great gusto. So if there is any PhD students here, there is hope. I mean, somebody might actually read the pages you slave on. It might take a pandemic, though. What I appreciated so deeply about her dissertation is how she describes in such a simple manner how these descriptions of utopia function as stereotypes and activate stereotypical patterns of imagination. They are composed of associations, disparate things we see as part of a whole, as if those associations are grouped together in our brains somehow. So they are sort of countries of the mind. We reuse these patterns to compare them to reality even though these places in all their perfection of sameness in all their cliches never existed of course culture as a subconscious in other words for better or for worse how immensely human and terrifying and beautiful this is. I mean, to me, this is why art, literature, cinema, theater, media art, culture, in other words, is my spiritual practice. It is a conversation with voices that have passed away a long time ago, and voices yet to be born. What we keep and canonize, collect, transmit as a culture, as humanity on the whole, is therefore so incredibly important since the biases by which we live are determined by this, by what we keep and transmit and therefore perpetuate. So in in trying to structure Sophie's countries of the mind, I came to the conclusion that the four utopias she proposes in her writings can be put on the extremities of a matrix of two opposing ideas. With on the y-axis, conservatism versus progressive thinking, and on the x-axis, a collectivist stance versus a more individualistic view on life. Both axes imply, of course, a spectrum, but as reductive as this seems, again, for better or for worse, the matrix is kind of a correct representation of what goes on in our minds while reading the books she quotes and while thinking of how utopia or dystopia should look. And here goes my reductive little mind. What I thought was ooh a matrix that's a perfect score system for a video game. Now I am not exactly a game designer. Just to give you a little bit of background of what I do, I sometimes make interactive installations like this one. Maybe some of you were here back in 2015 when I presented Temps Mort. I don't know. when I presented Temps mort, I don't know. It's a weird interactive short film activated by a phone call that the spectator makes with their cell phone. And when you call, the phone in the work rings and a very awkward conversation starts between people, these people, around the dinner table. The story is about a death in the family and family members hiding behind social decorum and their cell phones as not to talk about what actually happened. I sometimes make nonlinear interactive A to Z cinema films like this one. Could I have maybe the sound? I guess there is no sound coming from the HDMI. It's maybe in my own settings. Right. Sorry. I'll adjust for a second so that you can actually hear what is going on in the film, because otherwise it's a bit hard to guess. Where is my system preferences? Go ahead. It's on the other screen. Let me see. Sorry for this. No, there's no way of doing this. Is it black magic? Okay, I'll try this if it doesn't work. And this. Okay, I'll try this if it doesn't work. Nå er det en hel del av de flånige. The Vågåsjärden So, society is based on Franz Kafka's Gemeinschaft. It's a story about five friends and a sixth one that really wants to join them. And they don't know why, but they won't let the sixth one in and because of the outsider they have to actually question why they are grouped to begin with sometimes I make more educational installations like this AR activated building a public commission I did in collaboration with MVRDV and artenders on the occasion of the Floriade, a Dutch flower festival. For this installation, we actually mapped 1,800 native Dutch species of plants onto the four facades of a building, accompanied by an AR app that functions as an interactive map. It tells you where to find the plants depicted on the facade in the arboretum surrounding the building. So having said all this, I don't actually know if I fit into the category of animation expanded or cinema expanded, or if I just tell stories in an unconventional way, mostly involving some buttons. As someone at the digital theatre panel said this week, I think what goes for me most is that I like to blur the line between the spectator and the actor. So back to gaming. Anyone remember these? Yeah? Are there a lot of gamers in the room? I see some people of my generation, so I'm sure you had those. I had those, for sure. Texting as well. Remember that? With the buttons? So as many of you here, I grew up with a Nintendo, a 32-bit. Yes, I am that old. We got this little wonder of tech somewhere in the mid-90s, I think, and I played it as much as we were allowed to, even though we only had five games, because, you know, my parents were rightfully stingy. In the hours that we weren't allowed to play, we actually read books we found on the attic. And this is really a confession of nerdery right here. These. Does anybody know these? Yeah, you're too young to know those. So they were actually sort of a forerunner of video games. They were choose your own adventure books. And this might sound really boring, but it was like you would, there would be like the sentence, like if you're at the clearing in the forest and you want to go left, turn to page 140, to turn right, go to page 344, something like that. And the story would continue from there. You would usually start with a set of powers, which you could gain or lose. And there were kind of interactive books, which you could play. I mean, this was before flat screens, so, you know. Very nerdy. So I thought, wouldn't it be fantastic to make a video game based on those books? A choose your own adventure game, a bit like Bandersnatch, but then with a more philosophical and intellectual stance. So I made Autominem. In Autominem, you the spectator become the spectactor. This is a word I learned the day before yesterday. I'm very happy with this word. And you play Change. So Change returns to its hometown, looking for recognition for all it has done. And there is an event held in its honor, but after speaking to friends and foes, Change actually understands that it's never really welcome. So there are four levels in the game corresponding to the four utopia. The four characters are the four utopia embodied, so to speak. The world that they live in changes according to your answers. Do you agree with their stance? Then their environment looks more utopian. Do you disagree with them? Then the world behind them transforms in a more dystopian place. So just to give you a little bit of insight, this is kind of the script that is behind it. There are four different endings, depending on how progressive or conservative, individualistic or collectivist your answers are. I started out writing the script interwoven with quotes from the books that Sophie uses in her thesis. But they are actually mainly of interest to literature scholars and adjacent nerds just like me. So I got a bit stuck in the end. And what does one do in times of writer blocks these days? As any sensible person. I googled what to do writers block. Not the full sentence. I'm not a boomer. And the wormhole of procrastination hours later, I landed on some blog posts about GPT-3. So I don't know, does everybody here know what GPT-3 is? I suppose not. Okay. GPT-3 is a generative pre-trained transformer, which doesn't mean anything at all. It's a deep learning network, or to say it in popular terms, it's an AI. What GPT-3 does best is it generates text. So GPT-3 has 179 billion machine learning parameters. That is a lot. It is mainly trained on a data set of text available on the web, as you can see, but also, as you can see, on 67 billion books and on Wikipedia. So it works like a chatbot. You can use it, well, I mean, you can use it in several ways, let's put it that way. Depending on the software or the client you use for it, you can either feed it text and ask it to complete, or you can ask it questions and enter it into very, very strange conversations, or you can ask it questions and enter into very, very strange conversations. Or you can treat it like a character. I mean, like some software clients that use GPT-3, they allow for you to set a mood or a political stance, which I found very interesting. And I thought, aha, this is a good idea. It triggered with me the idea to leave the nerdy literary sources a bit behind and dig deeper within the history and philosophy to see how ideas on conservatism, individualism, progressive thinking, and collectivism have been translated within the political realm. And then, of course, have been interpreted by the unconscious of the internet, GPT-3. So finally, I decided to use GPT-3 as a sort of a co-writer. So having it compose some of the fiction, and like helping me to say, to shape each character's political mood. It also helped me from time to time to do some research about political quotes or whatever. Most of that was my individual human work, to be honest. I mean, I read a lot of philosophy and sociology. It was a good excuse to just put my nose back in the books. So let me just show you the opening scene of the game as an example, because this is all very abstract. Eindelijk. Eindelijk. Gezijter. Het is een verandering. Het is er. Eindelijk. Ik ben precies de enige die er staat om te wachten. Welkom. Welkom. Of moet ik zeggen welkom terug? We hebben zoveel... Ik heb zoveel verwachting. Words rise to the surface of your deliriously tired brain. Her language is vaguely familiar to you. Speaking it feels like a cold stone in your mouth. Not unpleasant, but not quite familiar neither. But language is not a mere artifact in a cabinet of curiosities, or a museum exhibit. It does have a definite function, in all human life. You cannot govern without it, there is no reality without it, there is no perception without it. The written idiom is not only a prime factor, but certainly one of the most potent, progressively so. Language alone can riddle and cut through the meshes. Used to conceal meaning, used to blur meaning, to produce the complete and utter inferno of the past century and centuries to come. Solely a care for language language for accurate registration by language avails. Your mind is rambling. The town might not look like what you expected it to, but at least someone recognizes you. So this first quote of the opening scene for example is from the collection of essays titled Make It New by Ezra Pound, writings that formed the foundation for modernism. And I found the source actually in conversation with GPT-3. It was most likely quoting someone's online peer-reviewed article on the failed utopia of modernist architecture, I think. Another comfort to you PhD students, you know, if nobody reads your PhD then at least an AI will crawl it from time to time. But as addressed widely in the blog posts that were, where I initially found out about GPT-3, there is also an enormous danger to artificial part of its intelligence. I mean, yes, it's like conversing with the subconscious of the internet in a way. And so, there is a lot of bias to it, since the data set on which it is trained is very, very limited. In the first place, to Anglo-Saxon sources, which, you know, as we all probably know, are not very devoid of many, many conspiracy theories, racism, sexism, in good, bad, or worse, English, ugly English, maybe, it is not a neutral source at all. And so you can't really leave it to do what it wants. It generates a lot of clichés, as do the style figures of Utopia, of course. This is exactly what Sophie's thesis addresses implicitly. These cultural tropes become frameworks to regard reality with, and they perpetuate a specific reality as well. Her countries of the mind reductively reflect on how we think about things in a reductive way. And I have to admit this also confronted me with my own limited data source. As a writing partner to GPT-3, my own intelligence turned out to be quite artificial as well. Since all the quotes I readily came up with myself, all the sources I readily consulted myself while digging through books of modernist architecture and futurist art and like serious serious German philosophy and the origins of hedonist thoughts were unfortunately mainly from Western old and often dead cis men. As an intersectional feminist I did queer or mainly femme the framework somewhat. And I say somewhat. I mean, I have to admit, again, the ready sources I had in the countries of my mind weren't intersectional at all. As for at Hinem though, in the end I felt that, in order to be true to the concept of the reductive matrix, I needed to express the clichés as they stand, in order to address the reductiveness. Now, spoiler alert. So for those of you making it to the end of the game, you can't win. I mean, there is no right answer to any of the questions. All they are meant to do is open your mind and your eyes to the limitations of what we think we know, where we think we need to go in order to achieve utopia. Neither is there any blame that awaits you there. On the contrary, there is something terrifyingly human about all of our artificial intelligence, actually. We are all trained on a very limited data source, unfortunately. The end song is a polyphonic hymn in acknowledgement of our limited foundations. hymn in acknowledgement of our limited foundations. Acknowledging this is key to subsequently diversifying our input, which might actually bring us closer to a healthier and more utopian intellectual biotope, I think. So, to wrap it up, what are the takeaways from this presentation? Well, as a famous white man once said, I know that I know nothing. Diversify your data source with more transgressive and less canonized sources. And don't worry, you're not working yourself into burnout for nothing. The next pandemic will surely make your PhD into someone's bedtime reading. Thank you. thank you so i think we have time for a little q a thanks a lot for this very inspiring talk i would say. Thank you. I think about it a lot. I would ask, try to ask something that always boggles me with these apocalyptic dystopian movies because they do not fit to what it would be in the end really. I guess it would be declination and extinction and these are two words that not really fit into a 90 minute movie or something like that. So is it the problem for storytelling or game making that things take very, very, very, very long in reality? I don't know. I think cinema is like a collection of sublimated moments or something. It's like an accordion of reality, right? Like moments that are too intense to actually live through, but we put them in the safe space on the screen and then they start to mean something. So I understand that like in the cliche of a blockbuster and an apocalypse needs to be chasey and whatever but I think there's other ways of talking about it to be honest ways that are more true to reality and that so because the gap between reality and the cinema what is actually seen on the cinema screen is less wide people will actually start thinking that this is a possibility instead of seeing it in these big Hollywood-esque ways. I mean, Children of Men, I don't know if anybody has seen that movie. I think that's a pretty good movie about the apocalypse. I mean, there's still a chase, there's still like a time element, but it sort of addresses the more human level of what it actually is about. To extinct. Yeah. Because I think that there are very popular ways in movies to argue problematic situations. But the problem is that if you have this kind of mindset in movie shows, you don't grasp the problem at all. So this is something maybe, yeah, this also leads to this bias anyhow. Are there questions in the audience? Right here. Yeah, I have a question, especially about the use of AI in your case. It's a rather unoriginal question. Usually people are arguing that they might go out of business or don't have a job because AI does the job so much better, for example with Journey or Dali for visuals. Do you think that with the use of GPT-3, AI could take over writing, I guess, as a book form yourself? Or do you think it could evolve into a valuable tool for writers? I mean, it's absolutely not going to eradicate writing. On the contrary. I mean, to me... I don't know. Okay, the introduction of lenses, for example, in painting in the 16th century, I think. I mean, when people saw these devices, the usage of the camera obscura, they also cried like the end of painting and whatever. And what happened was an immense revolution of an era of the most beautiful paintings on the planet. So I think if technology evolves, then it pushes us forward to evolve with it and to become even more creative as humans and to find our humanity contrasted to that technology. That said, I do think that there's parts of automation that are actually costing people livelihoods and jobs and that there is definitely a further centralization of power when it comes to that, which I find a bit frightening. And I think legislation is behind on that. But yeah, I don't think that concept artists are going to be out of business anytime soon. I think it's just like photography. Photography didn't eradicate painting, neither did it. So on the contrary, it made painting more interesting since painting didn't have to relate itself anymore to reality. So, voila. Any more questions in the audience? Here. I will give you my answer. Yeah, simple question. Could you talk a little bit about how you choose your visual styles in the video game? I don't know. I think there is a bit of artificial intelligence being steeped in Flemish painting my whole life in there. So I think that's where it comes from. I think, yeah, I've looked at one too many paintings in my life and then I come up with these weird painterly environments. Yeah, I think that's mainly the, I mean, it was hard. I have to say because the backgrounds, they are made in 3D programs. I'm guessing there are some 3D artists here. No? Yes,. So you guys probably know how hard it is to build something in 3D which doesn't actually follow natural perspectives, at least in the standard software. So it was a bit difficult in the beginning to force the program to do something more expressive than just putting two perfect cubes in a perfect perspective onto one another. So, yeah. Voilà. So, here we have a question. I will go. Thank you. I have a quick question about the... I think you call yourself more of a narrative, or your background is more narrative and storytelling and clearly you chose the medium of a video game to to kind of encapsulate this um what was good i don't think we saw it but in the game how do you interact with the system is it dialogue choices or is it just branching out and in and was there ever a consideration of adding Potential more because there are narrative games that do a similar thing sometimes Yeah, and some of them introduce a little bit of a timing element where you actually have to react to something more quickly But I assume that it's more just a narrative Yeah, it is and it's it's very simple as I said, I'm not a programmer at all And it's very simple. As I said, I'm not a programmer at all. So you can basically choose between two answers. It's multiple choice. And I don't know, I found that pertinent because it also relates a bit to how we live in a binary digital space and how it relates to the matrix. So it suited me well that I didn't have to put in any more advanced things, to be honest. And what was your second question? Oh, time. Time, yeah. No, because the funny thing is, so yesterday I went to check if the work still works. You know, that's always the thing with a new media exhibition, like artists asking each other, was your work stable today? So I went to check if it still worked. And there were 20 people playing together. And this is exactly what I wanted. You know, that you don't play it alone, but you play it in a group of people, and it triggers a bit of embarrassment, which I do deliberately as well, because there is a pop screen in it that at a certain point asks if you're actually really sure you agree with Ayn Rand. So it creates a bit of embarrassment, but it also opens up dialogue. I hope. So I expressly didn't want to put in a time element, because I wanted to give people the time to actually think about the questions. That said, there is a cut-off moment at eight minutes. It starts over. But if you can't choose between two things in eight minutes, then you're not a human. I mean. Yeah. So we could have one more question, I guess. Here comes the mic. Hi. So earlier this morning, one of the presentations talked about the difficulty of showing artist games. And so obviously, you've got it on display here in the Cyber Arts exhibition but I wondered where else you'd be showing it and how you're going to distribute it That is a very good question and that's still, it's a big we need to do some or I need to do together with my fantastic producer Melissa we need to do some research around this. Technically, the game can easily run online. And I think it's more of a game to present in exhibition spaces or in art spaces, because it does take a little bit of time and it's not a triple X thing, which anybody could just play on their computer. But I also see it in film festivals. Most film festivals have an AR, XR, VR department right now. So that would be fun. Yeah, so the thing is that it's technically very versatile. It's all HTML5, so once it runs, it runs solid as a rock on anything, including the internet in principle. So we don't, I don't know yet. If you have suggestions, please tell me. Okay. Yeah. Thanks a lot for your talk and answering these questions. Thank you. Alex Roerst. Thank you. OK. Yeah, thanks a lot for your talk and answering these questions. Alex rehearsed. ALEXIS MOUSSINE- Thank you. They will switch us on as soon as... One, two, yeah. So, I guess we are right on time. A little bumper here and there. So, shall I talk or do we send in the bumper? No, so I talk. Okay, so next up, Joni Goodman, The Road to Anne Frank. Joni Goodman is an experienced animation director and you know movies like Voids with Bashir, The Congress and the newest one, Where is Anne Frank? And he will talk about this eight year long challenging road working with 15 different animation studios and over 300 animators, right? So somewhere around this. So yeah, we are very glad that you are here. You could make it. Thanks a lot. Thank you. The stage is yours. Thank you. Thank you for having me. The stage is yours. Thank you. So hello everyone. I'm talking basically about the production of Where is Anne Frank? And this was indeed a very long production for us. Usually the films we did took about four years from script and development to the final film. And this one took a long time for various reasons, mostly budget reasons. And we start off as we always do, we start with development and when we started, we did Worlds with Bashir and it was a very small production, very low budgeted production which we did only in Israel with actually a rather inexperienced crew. So we had to find out a technique that works, that makes the whole production work and make inexperienced animators do the job and make the illustrations accurate. And we developed this cutout technique, which we used for the film, and it gave it a unique look. And in each production, we tried to do something different in the Congress, which was our second film. We started with sort of a hybrid between a cutout, the World Street Bashir cutout technique, and a 2D animation. And we didn't like it at all. This was like enough with this cutout. And we moved completely to 2D animation. And this required a lot of experience which we did not have in Israel, especially animators. So this was my first co-production experience. And it was a learning experience for me because I knew what I wanted. I knew how to manage my own team. But then you suddenly have to have several studios. And in the European co-production system, you have to get the studio where you get the money. So it ended up being with seven studios which was a lot and it was it was a learning it was like going back to university for me so it was a big learning experience and when we came to Anne Frank when we finished the congress I thought I now I know what I'm doing. This was the first time I said, okay, now how to run this production. And we started out working on the development. Can we see the screen? The computer screen? So we started out developing the character of Anne Frank. These are a few of the early sketches. These were done by David Polonsky, who did design for other productions as well. Fairly realistic, you'll see in the final product, this is still a rather realistic stage. One of the things we looked just to find something different, we decided to incorporate stop motion as background. And the development, this is like a few images of this is a concept art for stop motion background. This is a part of the fantasy element because there's a lot of Anne Frank's imagination coming into play, how she deals with her fears. And she uses what we call the army of light. It's like an army of mythological creatures and movie stars. So there's like Clark Gable and Athena and all that. And the biggest test we wanted to try out is the stop motion. So we did this development where it's going to open. It's going to open 10 times. Yes. One minute. Here we go. And the idea was to test how it works, compositing 2D with the stop motion. So we did a few... Do you think people love me just the same? This is a painting on a photograph. Well, that depends. That's one technique we tried. Later we have a live-sized room. And in the end we have the radio, which is completely done stop motion. Hurry downstairs. The broadcast is starting. So this is like an actual womb. Full sized. But when we go into the radio, this is like an actual womb, full-sized. But when they go into the radio, there's a stop motion. the battle for Leningrad and a million more of the city's residents, some one-third of its civilian population in all, died of starvation. But tonight, as we greet the new year, 1944, the battle moves from occupied Europe's strength to the heart of Germany. Within weeks, Allied forces will take Berlin, the seat of the Third Reich. I'm sending love to private Erzsie Wachstum. Sending love to private and devout... The spirit of freedom has found a safe and abiding home. To the Netherlands and to all other victims of German aggression, beware. And the fantasy element which is completely drawn. So we had this development, and while we were working, we did all the stop motion was done at Andy Gant's studio, who also did the, he does all the, I got the name blocked now, the fantastic Mr. Fox and Isle of Dogs, he did all those movies, an amazing artist, Wes Anderson, sorry, just popped out of my head. Wes Anderson, yeah, sorry, just popped out of my head. So Ari Foreman, the director, he liked it so much, he said, I want to try another set, just I want to try it completely doing it in stop motion. And David Polonsky really liked the idea. I actually didn't, because I didn't think it was right for the film. Plus, my agenda was, I finally know how to do a 2D production, so let me do it. But he did this test anyway. This is just a few of the puppets he used. An amazing piece of work. Like they did, they created a whole, they just created one scene. To me this is something I constantly tell Ali when we have screenings, that to me this is an example of why it doesn't work, because it's brilliantly done, but there's something missing in the character. And again, brilliant animator did that, but I felt like it wasn't enough for the story. Not that I don't like stop motion, I love stop motion, but Not that I don't like stop motion, I love stop motion, but it just didn't feel right for me. So at this point, we went into funding. And funding took a long time. Funding took like two, maybe two and a half years. And during that time, we were commissioned by the Anne Frank Fund to do a graphic novel. You might have seen it. About Anne Frank, not about the movie, based on the Anne Frank diary. And in this graphic novel, I did the storyboards. David did the final illustration. Ari wrote. And we worked on it for like 10 months, I think. And what it gave us, it gave us something we didn't have because we had time to think. That's one good thing about animation productions. You have a lot of time to think about stuff because it takes so long to create. And one of the main things it brought is a little levity, which was missing. So something very grounded, very realistic about the original development. And this threw us in a different, more stylized place, like softer lines, softer designs. And it really influenced the script as well. There was originally just, have any of you seen the movie? Yeah, so not a lot have seen it, unfortunately. But the main idea of the movie is that when Anne Frank wrote her diary, she wrote it in letter form to an imaginary friend called Kitty. And what we do is we create Kitty as sort of an imaginary friend she has conversations with. And in a year from now, as the movie says, Kitty wakes up in the Anne Frank Museum, and she goes like a lightning storm, and she comes out of the diary into today's world, and she's looking for Anne Frank. And she doesn't know what happened, because Anne Frank didn't end with the Nazis capturing her. It ended like two weeks before she was captured. Just a normal entry in the diary. So for Kitty, Anne is still alive, and she doesn't know what happened. So the movie is intertwined between flashbacks of Kitty reading the diary and remembering Anne Frank's life during the family hiding in the annex. And today's world, and Kitty is looking at today's world and what have we learned from the lessons of the Holocaust, have we learned anything at all? And she's basically looking for a friend. Once we will pass the graphic novel, at this point, David decided he drew enough Anne Frank because of the graphic novel. So he was replaced by Lena Gooberman. We'll see her illustrations in a bit. And the first step in every production we do is the voice recording. So in some productions the actors would also act. If they were sitting in a restaurant they would actually sit next to a table. But here we just did a reading session of the script. And that session was edited by our editor, Nili Feder, and this is just an example of a scene from the movie with the actors. Anne, I'm here. Come down, Anne, I'm waiting for you. Why are you hiding? No one can see you're my imaginary friend. Anne? Who is that? Is he a boy from class? Of course. Who is it? I don't remember you telling me about him. Anne! Oh, that's Sam. Sam Solomon. He's madly in love with me. Come down, Anne! He's a tough guy, but he's a brat. She's in game of phones now. Well, he is certainly good looking. And then the next step is Ari and me, we sit down and we just talk about the edited scene and we start directing it, managing it into shots, like what's a talking head, what's an action scene based on the audio recording, and we create an animatic. There's a storyboard, but the storyboard is basically allowed for the animatic. Usually animatics are just storyboard pages on the timeline. Ours, and we got used to this way of working since was with Bashir, are a bit more elaborate. There's lots of acting, lots of sketchy motion. And the idea is we want to put as much into the animatic as we can so when we get to the productions, there's fewer mistakes and fewer questions, so to speak. And as a director, I can look at the scene and say, this is working, this isn't working. And when I sent it, and this is another thing with the co-production, because I'm not sitting with the animators in the same room, they can't ask me questions as they are working, and hopefully the animatic gives them as much information as they can about timing, about acting, all the stuff that I can as a sketch. So this is the animatic for the same scene we just saw. Just so. Anne? Anne, I'm here. Come down, Anne. I'm waiting for you. Why are you hiding? No one can see you. You're my imaginary friend. Anne? Who is that? Is he a boy from class? Of course. Who is it? I don't remember you telling me about him. Anne! Oh, that's Sam. Sam Solomon. He's madly in love with me. Calm down, Anne. He's a tough guy, but he's a brat. Toughy. Well, he is certainly good-looking. Yes, but you'll see. Father will chase him away in a minute. Sam? What are you doing here in the middle of the night, for goodness sake? Don't you have a home? It's dangerous to be out this late. You know, if I were you, I'd go see him right away. Even now, in the middle of the night. Well, you know, he's not the only one. What's that supposed to mean? Up until a year ago, everyone was in love with me. Everyone? Of course, everyone. All the boys in class. Well, maybe not everyone, but let's see. Rob Cohen was in love with me since first grade, but I can't stand him. He's a baby, a hypocrite, a liar, and a crybaby. So you have this scene, and then... Just to show you the final bits, what it looks like when it's over. This is what the scene looks like. So one of the things we changed in the production, initially we thought that all the backgrounds would be stop motion. One of the main concepts that we changed is that the stop motion is restricted to the annex, to where they hid. And it worked because it gave us this claustrophobic feeling of something different, like a different universe. When the world outside, especially Anne's memories of life before the rise of the Nazis, is very bright and very sunny and very alive. I'm going to show you just a bit of the animatic of the scene I want to talk about, and that does involve stop motion. Excuse me. I have a girl with a high fever here. She's delirious. Clear the way. Clear the way. Clear the way! Thank you. Anne! Anne. Just a second, Mother. Anne, what's taking you so long? Everybody's been waiting for hours. Come downstairs. Why don't you go down? I can't stand her. Your mother? Mother, too, of course. But she's not my biggest headache right now. I told you, it's Auguste Van Damme, or Madame as I call her. She just sits around in her fat derriere feeling sorry for herself all day long. You should have seen them when they arrived last week. The minute they walked in, I took one look at what they'd chosen to bring into hiding, and I realised what kind of things are important to these people. If I have to die here, I shall die like a gentleman. Peter, come down immediately. I can't. I'm dying. Peter never comes down. He's always up there in the attic, dying. It could be a heart attack, it could be lumbago, could be pancreatic cancer. You never know what he's dying of. He just continues to die. And Madame, covered with her fur coat, she thinks she's on stage all the time. Herman, I tell you, I shall die of malnutrition in here. My days are numbered. But if I have to die, I shall die as a lady. But Anne, maybe Madame really is hungry. Oh, please, we're all hungry. I know, but I'm... So, that's the scene. Now, this scene has to be designed, has to have layouts, and I think for the entire scene, now this scene has to be designed, has to have layouts and I think for the entire scene maybe five studios worked on it. So the prep work is very, very important and it starts with the design and the character design. For instance, Anne had a very big change since the development time. These are just a few of these are just references. We don't have a lot. And the idea is how to create a essence and bring it back to life. And especially keep a spirit, because there's and bring it back to life. Especially keep her spirit, because there's something very energetic and very alive and very intelligent about the way she looks. And these are very early sketches of Lena. She did most of them in a taxi on the way to work. Just a bit of searching for the right type of characterization. This is a bit more elaborate. And these are starting to get close. And eventually also how to work with Kitty, how to design Kitty. And she's supposed to be redhead, a bit more rebellious. And she's sort of her alter ego. But her facial features are almost identical to Anne. If you put one next to the others, they're almost the same, the proportions. These are very early sketches. And one of the final touches Lena did was just make them a bit taller. And they suddenly became different characters. Like this is much slimmer. And this also dictates a lot of the way the animation works. Because suddenly we could break away from complete realism and go into more flowing animation. And we have... Just to show, I think this is the most characters we ever did in a production. There's no need to change back to me because I'm going to show stuff all the time here. So just a few of the characters. These are the early designs of the family. They changed into this. And this is the lineup of most of the characters in the movies are more. I just couldn't, you know, it's just a lot of characters from a, and the difference between characters from the past and characters from nowadays. And of course the fantasy which we saw a bit of in the development. This is actually the part with the horses is stayed almost identical in the movie. But a lot of changes were done to the background. But the animation itself was kept. And for Anne, we also had to design a complete Bible. Everything necessary for the animators. Again, these are all conclusions from the Congress. The more elaborate the design is, the less questions are asked. Because one of my biggest problems in the Congress was I had to constantly fix the design. Sorry. So the more explanations the animator is given, the easier it is for him to keep the model. And because we were spread out on so many studios, it was vital. So this is one of the early poses for Anne. This grew. Actually, the animators from other studios took it upon themselves. Every time they had a good pose, they placed it on this image and they sent it to us so we could spread it around, keeping consistency. This is the turnaround of Anne's head. image and they sent it to us so we could spread it around, keeping consistency. This is the turnaround of Anne's head. Each of the main characters had that level of turnaround, so an animator can really understand on each pose how the character should look in terms of proportion. Usually in these poses it's very difficult to catch the character properly. Vows, each of the main characters had the vows, different types of clothes. For Kitty as well, Kitty has three outfits she wears throughout the movie and other outfits as well in one scene. This is Kitty's turn around and turn around with a hat. So a lot of work was put into that part of the production just to get all the rules straight. Like just so each animator with the animatic and with the design could do the work from a distance. A few other turnarounds. This is Augusta Van Daan. She's the mother of the other family. Clark Gable. Herman. And an image you can see. And this is Margot. The backgrounds were in the, wait, where is it? Backgrounds were done, there's basically two timelines, one in the present, and the present is more water-colored and a bit monochrome. The pasts, these are just a few examples of line and then paint. And this is the present time and the past is much brighter. It's more, these are a few references. It's without contours. It's warmer. It's warmer, it's richer. And it's something that we both, like the three of us, Ari, Lena, and me, and we run, each of us, like Lena runs the design department, I run the animation department. And Ari's in charge of everything. And for me, it was like the... Today it feels a bit scarier, in a way. Like there's something warm and there's something that also relates to Anne's imagination and richness in the past and the future is a bit scarier, is a bit darker. And today, the present day. And I said something similar to Lena when I asked her, I actually asked her when I prepared this lecture, she said something, and this is the way she thought about it all the time. She says, the past is like it's all set. Everything is done, everything is complete, and the present is still not written. So it's watercolors and it's very loose. So just, it's like three answers that work, but we didn't discuss it until the production was done. And the stop motion, again, like I said, we did, they did an amazing work at the Endegan Studio. So just a few examples of the rooms. This is Lena's concept art. And this is her instructions. This is the detailed instructions of where everything should be. And they built mock-ups. This is something they did fairly quickly, just to get the proportions right with cardboard. And then they just started designing the props based on the illustration. This one, for instance, Lena said the bed sheet looks a bit like plasticine. It's not close enough to the illustration. They just airbrushed it a bit and suddenly it worked. This is just it was a lot of fun just choose your your tapestry and you can see the dogs from Isle of Dogs looking above. This is a few of Lena's comments. And this is the final room, based on the illustration she gave them. Another piece is the living room. This is the main room where they used to hang out most of the day. This, for instance, is a picture that Lena provided. I checked. It's nowhere in any shot in the movie, but it's there in the design. So every detail like this, there's no shot like this in the movie, but it's brilliant. So much work was put into it. And the idea of making the table very, very sharp, the lines of the table, to fit with the 2D animation. And making sure that the lighting, that everything works as a 3D environment, but still can work with a 2D character inside. That was one of the main goals. And this, when they took a photo, and they just had a session, they took photos. Some places, very limited. Some places, they had camera movements, which were a bit more complex. Most of the backgrounds are still shots. And they would take one shot with the characters in place just for proportions. The only problem is the assistant cut the behind too low, so they're like 30 centimeters lower than they should be. But this is another one, the family entering the room. And the attic, which is also an incredible set piece. And it's amazing how lighting changes everything. And these are all the sets, and this is the production dog. So the sets were like in boxes, and they could remove the walls and film from different angles. Or from above it was all pre-planned, of course. So the backgrounds were created in stop motion and then came time for the animation. And before the animation, I had a layout crew. And this is something that I kept next to me, like an Israeli crew. And the European studios were a bit skeptical if they could pull it off. But I think they pulled it off quite well. And the idea is that they give as much information as they can. It's not just characters on a shot deal with it, which sometimes animation production have that, where the layout is just placing a character on a shot. It's like putting as much information, as many poses, much of the emotion so the animator can actually start with the pose and work with it. And this is again all this is necessary because different animators in different parts of the world, I'm not there until I see animation. So as much information as I can, we try to provide as much as we could. This is Peter's mock-up and this is Peter's layout. Just a few poses. For the animations, this is again one of my conclusions from the Congress. In the Congress, I had a problem where the lines were too rough when they went into cleanup. The cleanup stage was done in the Philippines and in some other places, mainly in the Philippines. And the problem was that if the lines were not accurate, even if the line is a bit sketchy and it's like one centimeter to the left, suddenly there's distortions and I found myself in the end of the Congress fixing all the cleanup shots. So in order to do that, I asked for three stages. There's two stages of rough and then the tie-down. So the first rough is just my... And the studio managers hated that because that meant a lot of work. But for me, it was essential. The first rough is sort of my initial conversation with the animator. It does something very, very rough. It's not important to go into the model model it's not so important to go into specific details of the body just show me what you plan for this shot you have the animatic you have the layout you have the design show me what you plan and you could learn a lot from this this is really super rough and what's taking you so long? Everybody's been waiting for hours. Come downstairs. Why don't you go down? I can't stand her. But everything is here in terms of timing, in terms of acting. I told you. And these are fairly easy to fix. And in later stages, when you go into details and anatomy and all those areas where you need a lot of work, a correction might take a week. But here a correction can take just a day or so. Of course different animators have different drafts. If I have to die here, I shall die like a gentleman. Then the second stage is the second RAF pass, and here you start adding details. So same scene., lines are cleaner. Sometimes the lines are too clean. Some animators found it very difficult to do a rough line. This is again a stage for a rough line, but this is almost a clean line. And the red, I think, is one of my animators' corrections. Usually red lines mean corrections. We'll see a few of them in a minute. קורקציות, פשוט רדלינים מנסים קורקציות, אנחנו נראה על כך במינות. ארן. רק דקה, אמא. ארן, מה שמרחק את ככה? כל כך נסתם עוד עוד עודים. תצאי לבחור. הרדלינים הם שלי. אני לא יכולה להתגונן. אמא שלך? אמא, ככה, אבל היא לא בקרוב בי גדולה עכשיו. אז הנה, ההדה היא אנטומיה ומתק the idea is anatomy and accuracy, but it's not accuracy of the line yet, because the lines are still too rough for clean up. Like if I send this to clean up animators who are very talented but not necessarily familiar with the scene, they might mistake the proportions. But everything else is here. The acting is here, the models, the anatomy. If I have to die here, I shall die like a gentleman. And just a few examples of what the corrected. Every file that was sent in every stage was sent to me, and a lot of times I would fix it personally, or one of my teams would fix it in order to save time. If I could say it verbally, like do this, do that, it's fine, but sometimes it's just a matter of the proportions are wrong the arc of the motion is a bit off and we would prefer to draw on top in a in a red layer so so things would just move along. Otherwise, you could go into corrections that would span in weeks and even months, and shots get lost that way. Sometimes if you just send a verbal comment of fix this, fix that, fix this, fix that, it's going to the back of the line and you don't see it for a month. So a lot of time I would say, I fix this, move to the next stage. Don't stay on this stage, move it to the next one. This is an example. Anne. Just a second, Mother. Anne, what's taking you so long? Everybody's been waiting for hours. Come downstairs. Oh, another one. Thinking about it. She just sits around in her fat derriere feeling sorry for herself all day long. You should have seen them when they arrived last week. How am I on time? I don't have a lot of time. Do I? I'll just show you the final scene. Oh, just before that. This is, we wait with this. So after the complete tie down, what we call the line test. Come on. So this is the tie down. This is what's being sent to the cleanup. And it's a very clean line. But again, it's not... Can be a bit more accurate. And it's, very clean line. But again, it's not... can be a bit more accurate. And it's, of course, without in-betweens. But with this sort of line, the crew of the cleanup and paint, it's very rare that they miss stuff. Because the lines are so precise, it's good enough for everything. And... Because the lines are so precise, it's good enough for everything. She just sits around in her fat derriere feeling sorry for herself all day long. And the final in-between, and of course this is all without the in-betweens. It's just like the key poses. And the cleanup crew, they clean the key poses and then they do the in-betweens per the instructions of the animators. I don't know how many of you are familiar with that, but the charts, like, with numbers, this is like an explanation between frame one and frame five to another frame in the middle. So this is the final scene in in-between, just flowing animation. Anne! Just a second, Mother. Anne, what's taking you so long? Everybody's been waiting for hours. Come downstairs. Why don't you go down? I can't stand her. Your mother? Mother, too, of course, but she's not my biggest headache right now. I told you, it's Auguste Van Damme, or Madame, as I call her. She just sits around in her fat derriere feeling sorry for herself all day long. You should have seen them when they arrived last week. And... Just the final scene. I'm gonna skip ahead a bit. Excuse me, excuse me. That's it. Thank you. So maybe we can do a five minute Q&A if it's okay for Robert Seidel. I can see him right now ah here, is it okay? yeah so other questions in the audience otherwise I would start just with one I've seen Voices of Bashir like most people have I guess and I was quite impressed by this documentary style it has, and this very impressive way of flash it was at this time, I think. And then the Congress was even more stylized from the thing. But now, totally different thing. How did it come? How did you find, or how do you find this? What is the right style for the story that really supports the story? Well, we try to do things differently in each production. Just to try to test something else. A lot of times it's just like things that we found in one feature and we want to explore in the next. But mainly we just want to refresh. We're not necessarily interested in keeping our style, just finding something that fits the story. That was the main idea. And the stop motion was one of our triggers. And the graphic novel was, again, another trigger to say, oh, we can do this. That's more interesting to us. So we do the development. Most of the time we try to have the development look like the movie would, but a lot of times we'd say, no, let's scrap the entire development and do something else we found out in the end. So I read that you started as a journalist, am I right? No, no, illustrator. Okay, illustrator. Okay, sorry. I was wondering if there's a connection. But let me ask, you worked eight years on the whole production from the research. How important is this research, I guess, especially in this case, where you then took the advantage of magic realism, like it's called from some. But there's a lot of research you had to do. what is this way well this is mostly eyes work actually this is something that he would do and I always when he when he adapts a story to a film he would change it completely he did this I did the same thing with with the Congress which is based on the Stanislav Lem novel, but just keeping the essence, but making it the story Ali wants to tell, basically. OK. Is here a question anywhere? I can see it. I'm blinded. In the back. We got a question on the live stream. Thomas saying, that's awesome work. How many people are working on the set and what software do you use? And additionally, I love this style of art and I'm looking forward to see more of it. Me too. Thank you. We work on Toon Boom Harmony, mostly. And initially we started out, when we started out the when we finished the Congress we had seven studios because of the co-production system and we said that's hell, let's never do that again. So here eventually we ended up with 15 studios. Because I imagine my demands were high, and we had to keep up, and we had to get more money from different places. So eventually, I think the number is around 300 animators worldwide. So that's it. OK. So I guess no further more questions. There will be some, but afterwards, maybe. So thanks a lot for your talk. Thank you. Thank you. Thank you. Alles okay? So, yeah, welcome to our last talk of the day. In the last talk, Robert Seidel will talk about his work, his actual work. We'll talk about his work, his actual work. He's a Berlin-based artist. And we'll talk about kind of, he will explain much better, the overpainting of the world with a kind of biased imagery created by machines. Okay, so please welcome Robert Seidel. Robert Seidel. OK. So please welcome Robert Zahedl. ROBERT ZAHEDL, Hi. Yeah. The introduction is kind of equally confusing to my talk, I guess. And because I haven't been talking in front of an audience for quite some while, so forgive me and my sins. Yeah, I'm an artist. I do experimental animations on the edge. And... Nice. Mostly everything I'm doing is starting by a drawing. And the things I'm doing before the machine learning, I will skip briefly through it, because I think it has the potential to change a lot of things. Originally, I started out with experimental films. experimental films. These were my focus and that's the thing I graduated years and years ago at the Bauhaus University and being outside in the world doing animations or experimental animations it felt there isn't really a big interest in this kind of work. So my artistic practice expanded in the sense of the word and I explored more physical stages of the experimental image from media facets to architectural over paintings, which are specifically not projection mapping, but more loose and energetic and gestural approach to expanding animation. The music in the background, is it? Okay, that's okay and I'm also doing projections on self-made sculptures and kinetic structures and sometimes very unusual things like dinosaurs in Mongolia or dinosaur skeletons and of course the real ones are hard to get these days. I'm also working on natural surfaces in different scales from laser drawings to projection on vegetation and fog and water and though these things are kind of immersive, they're always rooted in my idea of abstraction. And then, yeah, to the kind of topic of today, I started to work on a new project, thinking about projecting drawings and manipulating them. And then the pandemic hit and all the things I've been doing so far just stopped or just got postponed. And it was really a very tough time to stay motivated and to get up at all. And then at the same time, kind of, the advances in machine learning started to grow to a point where I felt this is a potential field that needs to be explored and in this moment of being under challenged and oversaturated with divisions what could be done I found kind of a way to put the work I started into something with an aim and direction. So the first thing that really helped to transform was working first time on a film with an actual person, the performer Zuki. She's based in Berlin, coming from Australia. And she came to one of my exhibitions years ago and was sitting in front of my films and was really taking time to try to make something and to find an answer for herself. And after, I don't know, the second or third lockdown, we sat together and I brought my experiments along and I projected on her and she was improvising in her style that is maybe described to be between ballet, butoh and also the euphoria of Berlin club culture. So we both had these different layers of transformation, of art, of her body, of my confusion and we tried to layer them and create something intimate. And this was the kind of starting point of the actual film, Hysteresis. And while working on that, I started also to explore the possibilities these new ways of using machine learning and neural networks were giving to the world at that moment. were giving to the world at that moment. And it was interesting to see what was coming out of it and how the society that was trained into these could be channeled into maybe something different. So in the work process I shaped out something based on my work but disrupted through the machine and because in the beginning and this is still something that feels kind of engraved into most of the visual things in machine learning. After a while you start to look or you start to see a certain look and with kind of mixing my work with the machine work and with the performer, I tried to corrupt the images in a way and still maintain this beauty. And everything we did was, the projection we did in really just a small room of Zuki over several sessions, me projecting, me recording her, feeding things through digital processing, not only machine learning, that was too slow, but using TouchDesigner to really really build a system where we could improvise along and record these sessions and finally edit and editing them and dissolving them in these different possibilities of machine learning to create something very pulsating and full of twists and turns. And I will show the film at the end because now we are slowly approaching, or quickly, I don't know. My experiences with machine learning or artificial intelligence, but yeah, I mean, there were some mentions of it before today and and I still feel it's it's an thing that needs to be experienced really in the kind of dialogue with the machine because I felt it sometimes confusing and frightening. And everything I will say from this point is really just based on my experience and the experience of friends that are designers or artists over the last month. And I'm neither a developer nor a theorist. So maybe some of these things are a bit simplified, but I felt them in that moment. So I said I'm an artist. Some of my friends work in agencies and they choose that path because of stable income. But I felt in this time of research the split between artists and these commercial creators kind of started to fade because each sphere can simulate the style of the other's fear now and somehow the creation process somehow becomes secondary. You either take more conceptual approaches in the stage of making or what is the case often with the things you see in machine learning that it's just a random mixture of everything that exists and that's also something that hasn't been done really before these tools arrived. So yeah of course you know most of these things but but still I feel I want to say them. These machines sitting in tech companies and use of university laboratories or even just artist studios work through the history of mankind a creative history in in my case which is scraped from artists who are either dead and in the public domain or which is kind of need for discussion that shared their work just simply online for community or for portfolios. So now the work they put online and maybe generated a financial income in the way of kind of getting known is now threatened because all their creativity got harvested into something they neither got asked or paid for and it's interesting that that these companies are getting millions to to work in this field and potentially destroying jobs in different fields along the way. But yeah I'm quite skeptical because maybe a lot of people think oh that doesn't the art created by the machines at the moment isn't the real deal but I feel every month every week there's something new and it feels like there's a lot of things changing and when I started the film like more than a year ago and I finished it at the end of last year it's still or it already feels like something from a time capsule. So of course the discussion about copyright is clear. It feels like it's fully dissolving and the distinction between an original imitation or an inferior reproduction is eroding. And with this loss of origin there's also a kind of no responsibility and sometimes not a clear direction. So there's this primordial soup of creativity that can be shaped in any form without challenging the knowledge systems and hierarchies behind it. That can be good and bad, but of course there are obviously bad things like the bias that was discussed about several times, which is focused on a Western view on art. And me being born in Eastern Germany, I was also looking for art from the Eastern or former Eastern Bloc and this is also non-existent. So it's not just underrepresented regions of Africa, South America and Asia, but also the whole East is something that doesn't exist in a lot of these archives. So the idea that everyone will get access to creativity that is pushed by these corporations and communities is pushing marginalized positions even into non-existent sometimes because the creativity for the mass audience just is kind of pushing them out of the picture. So on the other hand the bias is pushed by especially corporations and in the last months there have been a lot of tools that not pushed by a community but companies like OpenAI or Google or in the Facebook domain. So there's a kind of censorship happening which is completely not transparent and when I started out working on the movie that was not a problem but if I put something into Dali these days sometimes it doesn't generate anything and it doesn't tell you what the word was and why it's kind of stopping the generation process and of course if you have a prompt you can kind of the bad word, but it's just the beginning. Everything they choose to take away, maybe because of their mindset, because of international rights and laws or just the lowest common denominator, is really completely unclear. So talking a bit about the creation, the idea of a prompt, I think most of you are familiar with it, is also something new because as an artist the last 20 years I was creating works that I felt couldn't really be described in words. So my task was to develop a visual language and now these prompts or text are used to help to shape my vision and this sounds good but after using it for some time you realize that your romantic idea of a proper description of a scene will work best isn't true because there are a lot of factors that shape the wording and that need certain triggers and now you can get some sheets or tools for prompt generation. So I feel this makes finding your own language even becoming harder. And when I started using, in my case it was VQGAN and CLIP, there was this system of negative prompts for example. I felt it was very powerful but in newer systems this is just kind of taken out because it would complicate the process but the creative power within the model itself is still there and it won't be accessible anymore which is an interesting development. As mentioned before the colonial quest of capturing all kinds of styles and types of data is very strong. For my work I was mostly using the very old WikiArts training set or data set which consists of 50,000 artworks by approximately 200 artists and of course you can train your own data but I found it interesting because I was working with my image what would happen if I'm using this standardized art historical latent space to use my images and my abstraction as a guideline and then there was Tsuki as a person and choosing specific styles also transformed her in a way so for example if I choose to create something in the style of a renaissance painting suddenly the muscles of Tsuki got emphasized and this was something that is logic in a way but still very surprising if you sit in front of it and see it for the first time, how Zucki is transformed into a painting. So these seeds... In the beginning of my journey, the images were created from random noise and then kind of iterated. And the beginning point for the film using this technique was using an initial image to infuse the process. And I even in the beginning had to create a little tool that was kind of putting all the pictures of my film into the machine learning process because I felt that's a good way to keep and maintain my own visual language. Some things that I found along this way is that the curation of the training data is not only relying on something around the bias but also about the quality. It took me, when I was experimenting with one thing, quite a while to figure out what happened because I was looking at something that was kind of similar to the prompt I created, but it had some strange artifacts and it took me a while to realize that this kind of pocket within the latent space was trained with badly compressed images so the machine not be able to differentiate between bad and good images just took it for granted that this style would look just took it for granted that this style would look kind of compressed. And for me, that was a big surprise. And of course, you see this in other cases, like because most of the training data consists of squares, is square shaped. It has a strong influence on the image composition, often heads are chopped off or the perspective is off-centered. And yeah, there are a lot of things to figure out along understanding what's possible and how to shape the process and the topic of curation is in that sense not only important for the data sets but also for the artists because of course you can be happy with the first image but it's I feel it's even harder to create something sometimes because you have these endless ways of changing things. The other thing, it also took me a while to fully accept is that we are not talking about image filters anymore or kind of standard image processing of... Sorry. Okay. Good, now I'm out. Good. So, yeah. I think that's it. I think that's it. I think that's it. I think that's it. I think that's it. I think that's it. I think that's it. I think that's it. I think that's it. I think that's it. I think that's it. Good, now I'm out. that we are not talking about standard image filters, but really something that understands a picture. So if we start, for example, giving the AI a straight line, it will make this kind of feel more realistic than you imagined or would be able to do with a standard filter in Photoshop, which existed before. And seeing this in motion makes it even more surprising and of course at the moment there's the problem of temporal incoherence that things are changing over time and introduced this flicker in details and there's a lot of research trying to stabilize that. But when I was doing hysteresis, all my images I put in were flickering and had these feedbacks where things are changing all the time. So luckily I could use this one thing that makes the AI harder to use in moving images. Yeah, I could make good use of it. And I think it's just a question of some months until this is fully smoothed out in a way. Yeah, the open source movement behind it, behind large parts, is also interesting because I was following the research in computer graphics for more than 20 years and for the first time it's possible that something that is researched and published in a paper is instantly accessible and with this endless number of tutorials and information floating around how to use these systems, it's accelerating creation even further. The thing that becomes more and more obvious in the next years is that if the output of an AI is problematic, the company offering it is potentially held responsible and not the person creating it. And that might be part of this bias and commercial thing but it's an interesting fact and a new reality for a tool because never in history before a tool was held responsible neither analog nor digital you could if you write an obscene pamphlet on a typewriter, nobody would be able to say something against the typewriter manufacturer or, I don't know, the person creating your pencils. But with the AI, there is this kind of branding that it's an all-knowing, in this specific field, entity and therefore it should follow a moral codex. I don't know, that was something that is kind of slowly coming and shaping, but I feel it interesting that the definition of a tool is also changing so strongly. The process from a not so technical but more emotional level is interesting because if you are working with it, it's easy to, of course, after you establish kind of how to use it, it becomes very easy to work in a flow state. So you are rushing to work on your image to find new variations and there's no friction because the system almost instantly generates something new. And I felt sometimes you are sitting there and you say, oh, I will stop to work on that. And this is just the last prompt I'm trying. And then some hours later you feel, okay, what happened? Because there is this potential in these tools that they become so easily and so addictive in their behavior that they really are in that sense also more than a tool. And I was talking to a friend who is an earlier adapter of Dali and he mentioned to me that sometimes in the beginning when it came out and everyone was like very hyped to get it, it came out and everyone was like very hyped to get it he would work without sleeping for several days and he told me that and that's one of the kind of anecdotes hidden in in my talk he told me that in the first weeks because the system then was free to use and now you have to pay, he would have spent $4,000 just for his kind of rush to try out things and to see what the machine would come up with. And this gamification of creativity is something I felt quite strong. quite strong. And this brings us to the question if this is also changing the idea of mastery. Because if there are specific AIs doing things, it will be sometimes very discouraging to learn a language or to work out things that are so easily to do otherwise now. But with this loss in actual work with your hand or within the process there's also quite an erosion within the idea, the concept and the refinement. And this detachment either increases the likeness of errors in comparison to a design language that has been established. Because if you're just using these tools that do everything for you, how can you fix things if it's not clear how to do it? So incoherences in images like these strange perspectives or penetrations in figures or deformed faces are there and they get solved, but if there's no solution on the technical level you might be stuck with the things you're doing. And yeah, the beginning of these visual tools, it was also interesting to see this push in obscure images because they were trained on fantasy illustrations or sci-fi images and video games and concept art. Sometimes these specific prompts to cite, for example, the Polish surrealist Bekszinski or comic artists like Möbius, were kind of pushing these obscure positions into a mainstream, so to say, in that moment. And it was interesting how the definition of art and a certain aesthetics is shifted by the biases of the datasets. And that's also something that happened in such a short amount of time that I was sometimes really out of a way to deal with it properly. And the other thing that also makes you wonder is we all know mass produced images. They existed since photography became available. But with the refinement of all these algorithms that exist in the moment, there's the chance that everything will lose any amateurish character, that computational photography and these machine learning tools in all their shapes and forms will push out kind of the naive image, the unpretty images and there's the chance that it will and in a certain field it already exists that you can create something with a click of a button which is a strange thing if you talked about something some years ago, how do you create your movies or how do you create your artworks. There was this assumption that there must be a software that creates it automatically and that wasn't the case and now it feels like this could happen. So what's on the horizon for art? It's clear that artists become a template for a future that is fully digitally assembled from fragments of the past. And that art history, in this case, I'm talking about painting, drawing, photography, of course not conceptual art or installations, but everything that is happening in a two-dimensional image or in a moving image, in a two-dimensional image or in a moving image, everything of their history will collapse into a single point in the present or the slight far future. And it will be really interesting that we get an easy access to an infinite reworking of the complete creative history. And the fact that the term art gets redefined by these technologies and companies It happened before, but here at least I feel that's the one revolution I will that was a big thing. And this, I feel at least, has the potential to be stronger and will create things that are beyond the user's level of cultural capital in a way. Two talks ago there was this feeling that maybe not everything will be replaced and I agree with that, but it's the question if you work in the commercial field, if the things you're doing and you're spending a lot of time on are worth for your client the money or if they just invest in something that is kind of generating something that is, even if it's just 95% of the thing you could achieve, it's still much cheaper in most of the cases in the end. cheaper in most of the cases in the end. So authors, musicians, designers, animators will have a hard time at least in certain fields if they're not highly specialized and have a very unique style. And even more precarious not so creative jobs like translation or voice artists, I think the impact will be even worse. And looking at it on a global scale, the click workers in India and other parts of Asia that label machine learning data or that build PowerPoint presentations for corporations or that are rotoscoping Hollywood film backgrounds, they will have a hard time because as soon as, for example, the rotoscoping is getting perfect, there will be hundreds and hundreds of drops just disappearing. And as said, there will be the need for something with personality and with uniqueness, uniqueness but because potentially these machines can learn from everything that we all did before it's it's the question if if there's still the room to to be unique and um yeah talking about everyone it's not just the creative the people in the creative industry, but also other people doing creative jobs like scientists. I mean there are algorithms at the moment that are harvesting papers for new drugs and new materials. materials and especially in these cases when it's about our health and maybe even the survival, it's very important that there is an ethical code implemented and being an artist, I often hear that art is useless and all this. And maybe it is to some people, but I feel these tools at the moment have the potential to open a discussion, a broader one, about what machines are capable. Because if people realise, about what machines are capable because if people realize oh they kind of can simulate a painting perfectly what could be the next thing so I'm almost through so the outlook the trust the manipulation of photography is nothing new, but I feel that AI has the potential to shatter any trust into authenticity. And with this very strong tendency to collect data, it's also interesting to think about if the AI knows our deepest desires and not just companies, how it is potentially able to manipulate us and to keep us entertained and within the loop forever. The fatigue is obvious because if possibilities are endless and access becoming easier and easier, there is the chance that we never finalize things and we never find something that really catches our interest. that really catches our interest. And coming back to the pandemic, I had days where I was looking on some streaming sites for a movie that should entertain me. And I spent more time just looking what movie I would pick than actually watching the movie. And I feel this overwhelmingness of choice is also a strong point within there. As for my film, I found it interesting and maybe this is just a very blurry thought but maybe worth refining that because there's the full access to human creativity, there's the potential to create alternative histories without the bias of the past and by revisioning these cultural terms through the lens of an AI maybe we gain a better understanding about ourselves and yeah I feel at the moment at least like there's no way back because even though they are challenging things like climate change and war. This is something that will stay. And I just had a discussion with a friend. I'm not sure if the process of creating an image is song whatever and and the energy you put in it and the learning curve and the pain of creating something not being really in pain but working through if this is lost i don't know what what the next generations will have to keep themselves into this idea of working on a vision. In the end I hope that real world experiences and interactions will stay important to us. And they were in the end the reason to finalize my movie. Because there's the chance that we all get stuck in these little loops somehow maybe. And yeah, I'm done. If you want to leave, you can leave, but I would like to show the film if there's enough time, because it will be shown tomorrow in the deep space as well, but I'm sure some people won't come, because it's kind of happening at the same time. I would say let's look to it because I totally could agree with everything you said because I'm also addicted to this stuff. So it was a very good, how to say, summarization of what's happening inside. But I think it's pretty good to see what's coming out of it, especially for people who are not addicted to the AI. So yeah, I would say thank you. Before we start the movie, maybe, thanks for the talk. Thanks. Vielen herzlichen Dank. Ich werde nur ganz kurz sagen, wir sind schon ein bisschen knapp an der Zeit. Um 6 Uhr wird das Museum leider geschlossen. Also eine große Q&A könnten wir jetzt auch nicht mehr machen. Außerdem glaube ich, dass es eher eine Q&Q wäre, weil die Fragen sind doch einige mehr als die Antworten, vermute ich jetzt mal. Aber vielen, vielen Dank für den Vortrag und vielen Dank für auch den Film. Wenn man ihn in 8K sehen will, wenn ich das richtig verstehe, dann ist es, oder 4K, dann am Sonntag um 12 ist ein Slot mit vier Experimental-Animation-Arbeiten, unter anderem eben Robert Seidl, Sebastian Pückner und noch zwei weitere, die mir jetzt gerade nicht mehr einfallen nach dem langen Tag, aber es ist auf jeden Fall super Eindruck, wenn man das im Deep Space sieht. Gut, wärst du jetzt noch hier, falls jemand eine Frage stellen würde? Okay, dann machen wir es vielleicht informell, dass wir hier Shutdown machen können. Vielen, vielen Dank nochmal. Ich danke Ihnen auch für den Besuch. Morgen geht es weiter. Wieder ein dichtes Programm. Wir freuen uns, wenn Sie wiederkommen. Schönen Abend noch. Thank you. I'm going to go get some food. Thank you. Thank you.