The video explores the intersection of art and artificial intelligence, questioning whether AI holds the potential to reshape the creative realm traditionally dominated by humans. The conversation delves into AI as a tool that artists can use to explore new creative possibilities and the broader philosophical questions it raises about the nature of creativity. Artist Matthew Ritchie shares insights into his experience working with AI and the evolving technology's impact on artistic expression.

Please remember to turn on the CC button to view the subtitles.

Key Vocabularies and Common Phrases:

1. ubiquity [juːˈbɪkwɪti] - (n.) - The state of being everywhere, omnipresent. - Synonyms: (omnipresence, pervasiveness, universality)

But surely, when it comes to the ubiquity, the facility, and dare I say, the sublimity of human creative expression, it's not particularly difficult to make the case for human exceptionalism, right?

2. sublimity [səˈblɪmɪti] - (n.) - The state or quality of being sublime, grand or elevated. - Synonyms: (grandeur, majesty, excellence)

But surely, when it comes to the ubiquity, the facility, and dare I say, the sublimity of human creative expression, it's not particularly difficult to make the case for human exceptionalism, right?

3. sacrosanct [ˈsækrəʊˌsæŋkt] - (adj.) - Regarded as too important or valuable to be interfered with. - Synonyms: (inviolable, sacred, invulnerable)

Will we find that AI establishes a kind of foothold in the once sacrosanct realm of creativity?

4. regurgitation [rɪˌɡɜːdʒɪˈteɪʃən] - (n.) - The act of bringing swallowed food back up to the mouth; metaphorically, a repetition of information or ideas without analysis or comprehension. - Synonyms: (repetition, reiteration, recycling)

Or will AI perhaps be permanently relegated to the rank of hack, right, merely a regurgitation algorithm that mashes up the real creative sparks that only humans can provide?

5. inference [ˈɪnfərəns] - (n.) - A conclusion reached on the basis of evidence and reasoning. - Synonyms: (deduction, conclusion, assumption)

There is a refining process, but the software itself, between the soup and your prompt is something called the inference fabric.

6. vernacular [vɚˈnækjəlɚ] - (n.) - The language or dialect spoken by the ordinary people in a particular country or region. - Synonyms: (colloquial language, regional dialect, common speech)

Can I populate them with anything? And this was the very early days of these generative networks.

7. verisimilitude [ˌvɛrɪˈsɪməˌlɪtjud] - (n.) - The appearance of being true or real. - Synonyms: (realism, authenticity, credibility)

It freed up artists to no longer have to try to have verisimilitude in their works and allowed for other modes of expression to take flight.

8. disjunction [dɪsˈdʒʌŋkʃən] - (n.) - A lack of correspondence or consistency. - Synonyms: (discrepancy, disparity, disconnect)

And a lot of new jobs will be made, and that economic disjunction.

9. gloopy [ˈgluːpi] - (adj.) - Thick and sticky; of a consistency resembling that of glue. - Synonyms: (viscous, sticky, gooey)

This was a kind of effort to kind of abstract that gloopy strangeness, kind of medieval psychedelia, I think.

10. subvert [səbˈvɜːt] - (v.) - To undermine the power and authority of an established system or institution. - Synonyms: (undermine, overthrow, destabilize)

The history of art is also a history of artists stealing technologies and then sort of subverting them to their own ends.

Visual Arts in the Age of AI

Good evening. Thank you. And thank you all for joining us for this exploration of the arts in the age of artificial intelligence. And look, we are not the only species on earth that creates art. But surely, when it comes to the ubiquity, the facility, and dare I say, the sublimity of human creative expression, it's not particularly difficult to make the case for human exceptionalism, right? I mean, from Beethoven to the creations of the human mind that soar above everything we see in the natural world. When you compare that to bowerbirds or bonobos, there's just no comparison between them, right? We don't see the flights of creative expression and the dramatic ways of expressing things that matter to us. We don't see that in other species, in the animal kingdom.

But as we head into what seemingly promises to be a future, right, where AI is going to be integrated ever more fully into our day to day lives, will we find that AI establishes a kind of foothold in the once sacrosanct realm of creativity? Or will AI perhaps be permanently relegated to the rank of hack, right, merely a regurgitation algorithm that mashes up the real creative sparks that only humans can provide? Or maybe will AI become a trusted assistant, allowing artists to explore realms that would otherwise remain uncharted? And perhaps maybe even one day, it will provide a powerful tool that we will look back on as remaking art as we know it.

Now, these are the very questions that many thinkers and artists are struggling with today. And the creator of the film that you just saw a moment ago, a fully AI created film, is one of the leading voices who has been pushing these ideas forward. And so it is great pleasure to welcome Matthew Ritchie, who is a visual artist whose work encompasses painting, sculpture, drawing, games, and video. His work can be found in the permanent collections of many leading museums, including the Museum of Modern Art, the Guggenheim Museum, and the Whitney. So please join me in welcoming Matthew Ritchie.

Good to see you. So that was an AI generated film that we just saw a month ago, is that correct? Yeah. I mean, when we say AI, it's a broad church, right? Of course, it sort of implies all sorts of frightening things, strange things that we've heard about in movies, some goblin like creature that lives in your computer. In this case, I'm using a kind of artificial intelligence we call machine learning, which is an enormous database that's been compiled in the last. Since 2022.

This technology is very new, and a small number of companies have done something quite remarkable, which is scrape. That's the word they use, the World Wide Web steal. Yeah, go ahead. They like the word scrape, I think. It's not that nice either. And then they've accumulated billions and billions of our images from YouTube, probably from your very shows, and I don't think that has made a big impact, but thank you anyway. And using that, they've sort of made a kind of algorithmic soup, and they use a number of programs and they use the soup to sort of regenerate this in a kind of Frankenstein like way.

And when I started working with it, the technology was based on what was called image to image, so you could smush together images in a way, the artist really, that's what we used to doing. It's like you smushed, but it's recently evolved. And one step further, something called a text prompt, which is much more like Siri or something, and you say, or you write in, make an image, and a thing emerges from the soup. You don't have any real control over it, and if you ask it again, a different thing will emerge from the soup. And this is the new strange, ghoulish material, which is also weirdly beautiful and fascinating.

You go in and no doubt stir the soup, right? I mean, you get some particular result. You like this, you like that, and you presumably try to form it through ever more refined prompts. There is a refining process, but the software itself, between the soup and your prompt is something called the inference fabric. And the inferential fabric is this. They call it a black box. Even the people that have made it, it's filled with all of the ideas that the programmers had about the material, and they've sort of labeled it all but pretty poorly. So I'm probably in there under mid successful artist or something like that.

That never comes up. It's not a high frequency search, but Taylor Swift's in there, I would imagine. So it kind of has a tendency over time to shift towards the biases of the programmers and the amounts of materials that are in it. So it's got a kind of power lore that drifts it over. So you're fighting that all the time, saying, I want to think about the films of Belatar, which are also in there, and so you have to kind of guide it so you're in a conversation, but it's a very strange one, like with the deaf mute.

Now, you asked a number of questions, and I would get them wrong if I don't read them. At the core of your short film for us was you asked, what will this new technology change? How will it affect what we believe is possible, what we believe is real, what we can do together, and who we believe is real. And I'd like to, by the end of our conversation, at least get the beginnings of answers to those very difficult questions.

But before we do, the question that always occurs to me in thinking about the power and the potential and the terror and the danger that people ascribe to this AI in the variety of ways that it influences life on planet Earth and will continue to do so every new technology over the course of human history, people have various fears and various optimistic outlooks of where it's going to take us. Are there relevant examples in the past? Is the camera a relevant example to thinking about the reactions then and compare them to the reactions today?

Yeah, there's a phrase sort of bandied around called technoshock, and it's how culture reacts to a new technology, usually with panic and fear at first. Then there's the it's against God's will gang, where like, this should never be allowed. And then usually the third phase is, I was always around. Who cares, right? We sort of adapt to things. It becomes so central to what we do that it's integrated and then it becomes impossible to legislate or moderate or go backwards.

You know, gun control is a good example. The guns hasn't been around that long. You'd think we would have figured out some sort of reasonable set of universal laws about management of that technology. We have no clue. Whoa, whoa, whoa. Now, when you say we, there are some places in the world, maybe american, we has not been able to figure it out, but there are places where people react differently and are able to figure things out. Well, I use that as a provocative example only to show how difficult it is. For example, in Meiji, Japan, they banned guns for 300 years. That's a long time to keep something out. But in the end, technology kind of finds.

So what we face is the second question, which is how do we adapt and mitigate the effects of technology? My favorite example is in Virgil's Georgics about 2000 years ago. He's complaining about the kids today. They've got these tablets, wax tablets, and it's going to ruin education because the kids, they don't have to remember anything anymore. Did they have a wax TikTok? Pretty much. I think it was like. So we've always had this kind of funny relationship, but the acceleration of technology is a different matter.

And I think that is much more to your point. It's around 1712 Jonathan Swiftwater Gulliver's travels, and he has the first example of GTP. It's a machine called the engine. It's a kind of big computer that generates nonsensical texts for the citizens of the society. And the reason it generates them is the citizens are now too lazy. They don't want to write their own texts anymore. So this sort of crazy steam powered machine produces new books. That's what he's observing at the very beginning of the modern industrial era, is our lack of capacity to sort of absorb technologies.

And that's a kind of question you're addressing in a way, as well. And so I think the, the dominant question when people think about art and AI is, can an AI system be creative? And there's so many ways of interpreting that question, obviously, as someone who's actually working with AI. And we want to show a clip of yours in just a moment. But where do you see AI on the spectrum of creativity, a tool for our use? Is it something that ultimately can be creative in its own right? Where do you come down on that?

I wouldn't want to speak for every artist, because the history of art is also a history of artists stealing technologies and then sort of subverting them to their own ends. Photography is a great example. There was about 100 years where photography was not an art. You made photography, you were a photographer, and there was no art photography in museums. And it took many, many generations of artists sort of pushing through that to get into the realm of like, oh, we think this might actually be useful for art. So at the moment, I'm very much of the church, that, well, we're going to get around to it eventually.

And being artists, we're, like, busy. We have nothing else to do, so we're going to find a way to subvert this technology. So let's have at it. But with photography, it's an interesting thing because some have said, and I have not studied the subject, and I presume that you have a glancing nose, you know, intuition for this, that the advent of the camera, while itself was not an art form initially, it freed up artists to no longer have to try to have verisimilitude in their works and allowed for other modes of expression to take flight. Is that true? I think it's for every portrait painter.

In 1860, it was sort of like, the jig is up, mate. Yeah, exactly. So that wasn't like a huge win, but it's like, oh, look at all the free time you have now. Yeah. So now you can go out and paint the haystacks or something. It does create an opportunity, but like, all technologies, they're hugely disruptive, and I wouldn't want to minimize the human cost of any changeover. And we're in one of those moments wherever a lot of jobs are going to be lost or displaced and a lot of new jobs will be made, and that economic disjunction. But to be honest, most artists don't make that much money. Is that right? Yeah, sadly, not. Like scientists, we really hit the jackpot banking it. You come up with that equation and you're set for life.

I want to get to a clip of one of your pieces. Color confinement. Well, this is a very strange piece, right? It was made during COVID And so I was an artist in residence at MIT, and suddenly MIT was closed, and I filmed with a gopro, all these empty spaces of MIT at night. And I was like, I sort of have the film sets. Can I populate them with anything? And this was the very early days of these generative networks. So I thought maybe what's living in MIT at night in this empty campus is the intelligence of MIT. It's sort of this sort of, like, reverberating around the spaces in some way.

Cool science ghost. Yeah. And at the same time, I've always believed that we have the ability, one of the great contributions of the sciences is produced a new language for art to talk about, like, new subjects, which is an amazing thing. So maybe the characters inside this campus are embodying the forces that all these scientists at MIT, they're sort of dreams of science. All right, let's take a quick look at that. On the red road it's 1976 the bases are loaded on the red road a blood maiden is ruined from the star it's tough to win two in a row on the yellow row it's 1976, give that bow chase I mercy I'm a magician I'm a new man here guided by invisible hands to an unknown end on a white road it's 1976.

Mirror cemetery on the white road, it's 1976. Oh, hell is thick with tests and heaps and piles of tests. A dark that house so white, a spiritual light, a darkened house, a white dagger, spherical mind. So it's a beautiful piece. Can you give us a sense what was AI's role in creating that? Well, it helped write the lyrics. Chat GPT-2 was a very, very funny. It just came out and it couldn't really make us a sentence. But I thought, oh, Emily Dickinson, you know, like, this is, it's a natural poet. And so everything it did kind of came out wrong. But I was like, perfect for me. So.

And then the kind of very abstract drawings that you saw, that was the AI doing image to image comparison, that were kind of those avatars from a video game wandering around whose sort of isolation suits. That's all from AI generated. So I use it. And the beginning of editing software essentially acts as a randomizer in a positive way. So it increases the level of chance and things that surprise you. Often one imagines that the goal of the development of AI in terms of the AI systems themselves, is to make them better, to get them ever closer to the kinds of things that we would anticipate, the good sentences and the insightful ones.

Is it the case that a better AI might have been less artful? Absolutely. I mean, imagine the best book. If someone came up with an idea for a novel, like the best novel ever written, and it was like the most realistic description of us sitting here. Like, totally realistic in every way. It would be the worst novel ever written. Not that we're not having it. Great. I hope you guys don't. Right. If it was just the sort of directive towards realism, given what we know about the strangeness of the universe and how I feel, it's not only is it deeply incomplete, it's incomplete aesthetically, it's incomplete philosophically, but it's also just a kind of incoherent directive, because realism towards what you've already pre decided, a kind of rather banal outcome.

If you say, I want to make a really realistic image of what. And one of my treats personally is, at the end of the day, when I've done some AI work, I'll type in something like the laws of physics to see what it kicks up as an image. It's usually like a Bentley because it's. Can you hear us? Yeah, it's been like, labeled as, like from all these car ads that have repeated and are like the laws of physics in motion. And it has no concept of anything remotely interesting. I thought it was because physicists make all that money that you made reference to, and so you made reference to science as providing a language, a new language that obviously we scientists use.

But artists have been inspired by the insights and the language of science itself too. I mean, has that played a part in work that you have done in the past? Yeah, I think it's an extraordinary contribution that we personally, I don't think we talk about enough. Is the change in visual language and cultural language. Science is a cultural technology as well. AI is just one small subset of it. Yeah, for sure. And science didn't always look the way it looked now. I mean, in 1750 or something, chemistry experiments were like a witch's cauldron. You know, that room was like Harry Potter world.

And so there's a visual language that comes with science, the notation, the descriptions, and a deep conceptual language that introduces scales and types of being that weren't really on the table except in mythology. So I think that's an enormous part of our culture we kind of tend to forget. Yeah, for sure. I mean, part of what these conversations are about is to really shift science to the place in culture where it can flourish most fully when it comes to AI, and potentially it giving rise to a new language. Do you have a sense for that language now? I mean, do you feel there's a creative quality that you couldn't have articulated without first immersing yourself in these systems? Yeah, I think there's a wonderful strangeness to the possibility that we might see in our lifetime a machine thinking and AI, what it produces is visuals, is sort of the closest we'll ever get to looking inside. Like the dream state of a computer, like following along with its processes.

When I make these films, I pull out the really most coherent and understandable parts. Most of it is just like a dream, like a Salvador Dali painting. Like the thing doesn't know anything about space or time or bodies. So it just like a rabbit grows out of a head and it's like, is that what you want? Kind of. Now I do, but like, one of the directions that when you speak to forefront leaders in AI, there's some of a debate right now in terms of how important the current developments will be in the long term for AI. You know, we've had conversations on the stage where leaders have said, a few years from now, large language model type technologies, it'll be gone.

Because the advance requires the very thing that would prevent the rabbit from coming out of the head, namely, give the system a model of the world so that it's not simply based on probabilities and enormous data sets and just figuring out some kind of mashup that naturally aligns with the prompt that you gave it. So if you had a system that had a world model in it, would your process then be radically different? Are you then dealing with more like an assistant who knows about the world but has access to all that scraped data that you can't hold in your head? Is that something that's appealing to you, and can you imagine the process changing?

Well, it's funny to hear you say that, because those are the same people that say, well, the large language models like Lyon five B, it's 5 billion images. They say it's too complicated, it's already baked. We can't get rid of it. We can't go backwards. So either that's true, in which case every problem in Lyon five B and all the software is baked in. From the beginning, it's like a kid. It's learned that eggs are poisonous and you can't now unteach this knowledge. Or that's not true. But in order to start what they call general intelligence, they're going to have to get rid of all of those models which currently they're telling us they can't get rid of. It's too expensive than no one.

So I think there's a good case to be made for a public intelligence model. A library and a national effort to sort of really decide what is it as a culture that we see as the commonly shared space, and to make sure that space is ethical and doesn't contain at the moment, it just contains so much garbage. And they're really trying to sell us on both things at once. Like, this garbage model is the only model you'll ever have, so you have to buy it from us now, but in the future, we won't have this garbage model like one. Make up your mind, guys. Yeah, for sure.

Now, when you see the output of AI systems, I mean, there's so much AI generated visual art. Do you recoil at a significant fraction of it, or are you like, hey, that's cool. This is the next step in producing visual art? Well, we've always had kitschev, and kitsch is what most people like. And I think the reason that material looks the way it does is it's built on video game chips and the engine is a video game engine, and the material is video games. And gamers have been enjoying this kind of soft, anime, kind of plasticky surrealism for the last 30 years. So of course the system's super good at making more soft, plasticky surrealist images.

That will change if we ask it to change. It won't change if that's the common denominator. It's going to shift more and more towards the garbage y kind of suit. I mean, so can you imagine sort of a push that's coming from artists? I mean, the problem is it's very expensive. But can you imagine a push to create a database that is not influenced by the kinds of things that we're all familiar with seeing that AI look goop? Yeah. Is that something that you can imagine is that happening? Is that a conversation? There's the beginnings of that conversation, and, you know, it's expensive, but it's not in the terms of kinds of money we're talking about, the United States government could easily afford to do it, you know, for the price of a few bombs.

Probably one, right? Quite possibly just the one. So in terms of even the Library of Congress thing, it's totally feasible to start again and take this very seriously if it is indeed a serious thing in the way that certain thinkers are talking about as a cultural all embracing threat. And then also the other part of that is making it open source so that everyone can learn and have access to it, and this becomes our common shared cultural heritage. So that's what I. And I'm working on a paper for this, for MIT with Humagupta and Caroline Jones, and this is what we're strongly advocating for.

Oh, is that right? Is a kind of human genome project for culture? Because, you know, they said that all these things they said about the human genome project is too expensive. It's got to be private. It's already started. No one can stop it. And then. Sure, yeah, absolutely.

So I want to get to another one of your works, an excerpt from infinite movement. But I wanted to just spend a moment on allowing you to give us a sense of what that is and also the particular process involved there, which I understand involves sort of the AI version of good and evil as it is incorporated in paradise lost, or, you know, things. So fill us in on that. Well, one of the, I'm full of little funny historical sidebars, originally from England. So we have this, like always a grab bag of weird old tales, but. So John Milton, who wrote Paradise Lost, also wrote a famous screed called the Areopagitica, which was the first argument for copyright because at the time, authors were ruled by something called the Stationers Guild, which was a private concern because the government found it too difficult to manage.

Sounds familiar. So there was a double argument happening in all of Milton's texts, was like, for freedom of authorial thought. And when these conversations, that was just the first example now of someone using AI and getting a co copyright with an AI. So we face a kind of interesting question. At what point are we going to recognize these programs as kind of legally autonomous and in a sense, that's investing our humanity into them the way we did with corporations?

And that's a mirror that's on us, right? When we really. We said corporations are people. That was our bad. This was not a smooth move. So within these programs, the earlier versions, there was something called a generator that generates old millions and millions of copies of image. And then there was something, a program called a discriminator. Can we just show a little image just to visualize? So I gather this is associated with the generator. Is that so? I make paintings and films and sculptures and installations and performances, because I sort of feel there's a bit of everything in all of those things. So this was a kind of effort to kind of abstract that gloopy strangeness, kind of medieval psychedelia, I think.

I think it captures it well. And then the other, I gather, is the discriminator that you made reference to here, which, you know, in my mind, is sort of spookier. Its job is to choose among the generator makes. And the discriminator goes, I like that. I don't like that. And that's within the AI system itself, that that process is happening over and over again, generate, discriminate, and so forth as you go forward. And then ultimately, it yields this peace. And let's just have a quick look at that. The first kiss, the octave light, first kissed. Jordan is a player. Overthrow. The weaver, overthrow. I chose a question on the fruit of it.

So this is clearly an opera that we created. And, again, can you give us a feel for AI's role in creating. Creating this work? So I want to give a shout out to my collaborator, Shara Nova, who's the real musical talent behind this. So she wrote the music, and I wrote the libretto with the age of AI. And the idea was that these characters, the generator and the discriminator, fall into the. Well, we've all seen that cartoon with the little. There's the angel and the devil on the shoulder, and they're sort of arguing about what the universe is going to be, and various other characters show up who approximate kind of archetypes of our past. There's a kind of Eve like character called the weaver, which is a computational term.

They're all based on computational sort of thoughts. So it's like the early universe is beginning to think, and it's also beginning to argue over what direction in the CP mirror would like to go. And more or enough, the discriminator is always like, I think that's about enough. And the generator's like, no, we need more universe. And as you can, thermodynamics and so forth, things get more and more disordered and chaotic as a result of that. So the AI was wonderful in this, because you can sort of feed it anything, really, and especially early chat GTP. I fed it sort of paradise lost, and it sort of didn't really know what when that phrase silenced. You mean you just fed it text? Yeah, just text, yeah. And that phrase silence in the play of time, it sort of came up with these wonderful, confusing phrases that really don't mean anything, but you sort of imagine that is the space of creation itself, where nothing quite means anything yet, but it's going.

And so is this a new kind of creativity then, that we're seeing where human and machine are kind of working together, or is this just sort of a better paintbrush? I mean, how to think about whether this is a game changer or just a little extra juice in the kinds of things that you used to do before. I think it's up to us. I mean, the idea that there's potentially an accessible library of all human culture that we can all get to eventually, and that we are generating a toolkit that would allow us to play with that meaningfully is an enormously wonderful and poetic and democratic idea.

Unfortunately, at the moment, it's run by a small number of monopolistic corporations who are going to give you the paintbrush that only paints you what they want you to paint. And that's not such a wonderful. That's the stationer's guild, that's back to total control, and that's definitely their game. They've sort of said it pretty openly, I think we have. There's always the opportunity to play the game again, though. So I'm an advocate because every technology that has attempted to be controlled has eventually slipped out. So why mess around? Let's just advocate for it and bring it into the public domain for sure.

I mean, that sounds utterly amazing. If every human had access to this archive, that would allow them to build upon all the creative things that our species has put forward in the past. I can't. I mean, it's just enormously exciting. But a question I have that obviously is the vital one that will continue to come up in all of these conversations, which is when you think about AI as this currently configured, where it's basically putting together, mashing together, and synthesizing, perhaps as a nicer word, things that humans have created in the past, one can take that to mean there's nothing creative happening there. It's just moving the pieces around in a different way.

But could it be that that's what we as so called creative beings actually do in some sense? That we are basically pattern recognizing entities that synthesize things that we know about from the past in ways that appear novel because that particular combination has ever been put forward before. Could that be what human creativity actually is at rock bottom? Well, that's probably above my pay grade. It's more your. More your field. But I would say that I feel that our understanding of our dimensional place in the universe is as yet so poorly articulated. We operate in so many more complex ways that even though. Oh, so there's twelve particles, right? So therefore, everything in this room is made of twelve particles. So that's it, right? We can all go home. It was wonderful. I can do that math. Yeah.

But then the different configurations, when you mash those particles up in different ways, yield the rich spectrum of things. That and stars and nuclear reactors and trees and tree frogs and everything it does. So it is a kind of. Over time, that conversation is both interesting and because it's so universally a condition, it's also kind of not interesting. Maybe AI is a very sophisticated numbers game. Maybe all human thought is a very sophisticated numbers game. I think what it might be interesting to think about is, can AI occupy a space in relation to human thought that allows us to see ourselves and in new and interesting ways, rather than just becoming another paintbrush, it becomes a kind of paintbrush that is a mirror.

And that is where, I mean, I always think about Frankenstein in the original text, Frank, you know, he's created the monster, but the monster's supposed to be beautiful. And the monster says to Frankenstein, am I not your Adam? You know, are you not my creator? Why do you reject me? AI is like that. If it's going to be anything good, it's because we help it to become good. And that's an if. But there is a chance. Exactly that there's an if. So what about this whole development terrifies you? Is there anything where you feel like maybe this is the wrong thing to be doing, either as an artist or, more generally, as a species?

Well, as an artist, there was a rather bleak period where everyone decided after the Renaissance that paintings should look a certain way. And they looked that way for about 200 years. And if you were an artist, that was it. You learned how to do that. That's a depressing thought of a homogeneity, of a banal imagery. But I think that the lesson we learned from that was like, don't do that again, because everyone got.

It's just like, you go and there's miles of those kind of brown paintings in museums, and nobody even wants to see those things again. But the other terrifying part that has to be acknowledged is the energy use of AI is enormous, and they've made, 15 billion of those images have now been made using AI imagery, and every one of them costs an enormous amount of power. So if we keep gobbling up to make not very good cartoon images, that's not a good plan either.

But I guess the other thing that I'm just to finish up here that I wonder if it causes you any angst at all. What if these systems get really, really good and they start at some point to produce things that you start to feel like? You know, like I feel when I look at an Albert Einstein paper, you start to feel like, I'm never going to do that, and therefore, why bother? I mean, can you imagine getting to a place where these systems can just make the creation of the art that has given your life, presumably a certain amount of meaning, that you just feel like, hey, it's doing it better than I can, so why bother?

I'm going to come down pretty solidly on the side of human meaning that we are the ones who provide that meaning. There was a time when all billboards were painted by hand, and sure, now they're all printed. AI is really a photo based medium, so it carries itself through the screen. And as an artist, I can make small films, but I can't compete with a Hollywood studio that might produce. So I don't make Marvel movies, and I never will. It's okay. I mean, unless they want to call me, then I'm totally available and ready to pitch in. Artists make great Sci-Fi movies, but there's a strangeness to being human.

And I think ultimately what we want from our art is that strangeness, that sense that another human made it. There'll be a place for AI art just like there's a place for Marvel movies, but it will never fulfill that space that we want of art, where a person showed up and made a beautiful, strange thing. It will displace lots of jobs, but that job always migrates to the next spot of the weird. Right?

We seek the weird for sure. I think the same true in string theory, where there's efforts to sort of explore all of it mathematically, and then there's another physicist saying, well, sure, now you've got just a bunch of math, right? But you don't. They don't know what they're talking about. It's a nice thing. Exactly. Well, this is a wonderful conversation. Matthew, thank you. Such a pleasure. Thank you for joining us. Thanks. Matthew Ritchie.

Artificial Intelligence, Innovation, Creativity, Visual Art, Technology, Matthew Ritchie, World Science Festival