The Hour is Blue

AI Art and Artists: A Conversation with Dr. Dominic McIver Lopes

In this conversation, I speak with Dr. Dominic McIver Lopes, a Fellow of the Royal Society of Canada and University Killam Professor at the University of British Columbia. In a dozen books and around eighty articles, he has offered new conceptions of aesthetic value, art, and the relationship between art and technology, especially new information technologies. He advocates for pluralism about artistic and aesthetic value and is currently writing a book about value pluralism in the global history of philosophy.

Our conversation centres on one recent article: “AI Art and Artists: What They Are, What They Could Be, What They Should Be,” published on March 8, 2026 in The Journal of Aesthetics and Art Criticism.

We discuss what AI art is, whether an AI system could ever qualify as an artist, and what questions about agency, creativity and value reveal about the distinction between making art and being an artist. The conversation also considers what AI could become, what kinds of artistic roles we might permit it to occupy, and what remains distinctively human in artistic practice.

CiTR 101.9 FM / The Blue Hour

Recorded live on October 6, 2026

Listen: CiTR | Apple Podcasts | Spotify | YouTube

The Blue Hour, hosted by Farha Guerrero, airs live every Tuesday at 2 p.m. PT on CiTR 101.9 FM at the University of British Columbia in Vancouver, Canada.

Transcript

Transcript lightly edited for clarity and readability while preserving the substance and natural rhythm of the conversation.

Farha: Welcome, Dom to The Blue Hour.

Dom: Thank you, Farha. I'm so happy to be here.

Farha: I'm going to get you to read one of the opening paragraphs from the article. But before we go there, I want to mention one image that kept recurring in my mind while I was reading it.

It was Edvard Munch's The Scream, because running through your paper is something you repeatedly call existential anxiety, existential dread, perhaps. That painting has always felt to me like it captures something of that.

So maybe we'll begin there. Why don't you surprise us and read the opening paragraph, or whichever passage you think works best, and then we'll start talking about it.

Dom: How about I read the first paragraph and the last one?

Farha: I like that. Let's do that.

Dom: Here's how the paper begins. By the way, it's open access, so you can all follow along if you like.

ChatGPT, Midjourney, and Udio generate stories, images, and songs that amaze us. They can trouble us too, provoking existential anxieties. Existential anxieties are anxieties about affirmative answers to questions like this. Are they art? Could there be AI artists? Would they be creative? Might they surpass us? Will they, as my friend Lindsay Brainerd asks, challenge the special status we've taken ourselves to have by busting us in the domains that we cherish most? Anxieties that drive us to philosophy often persist despite any assurances that philosophy might offer. And we'll no doubt continue to worry as long as AI technologies remain unfamiliar and uncanny. Philosophy can only settle whether we have reason to worry.

So the first section of this paper argues that the existential anxieties are not justified given the current state of the technology. The second section argues that we have no reason to worry about future possible AI and AI art and artists if they were to turn out to be just like us.

Well, maybe they won't be just like us. In that case, the final section proposes that we should undertake a normative inquiry where we ask what they should be. Such an inquiry replaces existential worries with practical ones. Therefore, no existential anxiety is warranted.

In the course of the argument, we can draw some lessons about AI art and artists and also about our values and interests.

So that's how we begin. And this is where we end up. I guess we'll have a conversation about what happens in the middle, the meat in the sandwich.

Perhaps existential anxiety is a symptom of enlightenment. Copernicus discovered that we are not the centre of the universe. Charles Darwin, that we are not enthroned atop the tree of life. And Sigmund Freud, that we are not masters of our own selves. Their discoveries disorient some, even today. Geoffrey Hinton might deserve credit for breaking up the human monopoly on art. We're not there yet, but we should welcome the event if AI technology presses us to identify our aesthetic interests and to advance them in ways we never expected.

Can I just say one thing? What I want this paper to do is say something about AI artists and the art that they make, if there are any. But more importantly, I want the paper to say something about us, as if they're like the mirror.

They're us in a mirror. Right now, the image in the mirror is not that flattering. But I want us to think about the possibility of more interesting and perhaps even flattering images.

So this is about us as much as it is about the technology.

Farha: Absolutely. And you certainly go deep into that, especially in part two.

Let's unpack existential anxiety, since you've mentioned it at the beginning and then you bring it back. What exactly makes existential anxiety? This is a deep question philosophically.

Why do we have this as humans? How can you break it down? I gave you an image of Munch's The Scream, but what is it really?

Dom: I don't want to say what it is. I think it's an intangible.

There are some phenomena that get us thinking about ourselves, thinking about the world, engaging in inquiry - scientific, philosophical, humanistic, social scientific, what have you. There are phenomena that get us thinking.

I say that these phenomena raise primitive questions. They're the questions that get us thinking in the first place.

I think it's very important for us, as we conduct our inquiries, to go back and check to see how well they're doing at addressing those primitive questions, existential anxiety, for example.

But I think it's also important that we remain open about what exactly the primitive question is asking, or what that phenomenon is. What exactly is existential anxiety?

I think we learn more through the processes of trying to figure out what it is than coming to a conclusion as to what it is.

We seek conclusions, determinate answers, but not to all questions. For the primitive questions, I think we should be willing to leave them open.

So I'm not exactly sure what existential anxiety is. I just want it to be there, prompting us to inquire into these matters.

Farha: Is it because it's a feeling that makes it intangible? Is it because we can't rationalize it?

Dom: We can rationalize it, but it's probably not prudent to rationalize it. If we're rationalizing, we're cutting it down to size, and then we're missing an opportunity for it to motivate future inquiries.

It should be there continually bugging us. It prompts us to try to understand ourselves.

We get some understanding. That understanding is valid. We say, now I understand myself better.

But the primitive question remains, and it's continuing to motivate new inquiries so that we're moving forward.

Farha: It's like a catalyst in a way.

Dom: It's the initial impetus, the initial push.

Farha: You start off making us feel that anxiety, and now you're trying to unpack it and see how it relates to how human beings are seeing AI art. Why is it making us anxious? What is underneath this anxiety?

Certainly, as we discussed before coming on air, this is a very topical discussion, happening in real time, and our reactions to AI art are also happening in real time because we are being exposed to it more and more.

The next bridge, is this idea of agency and intention. How can we understand AI as having intention when there's always a prompt and something that we are initiating? Can we start there?

Dom: This is just looking at how things stand in the world right now, what the technology looks like, which is quite different from six months ago when I wrote this paper, or a year ago when I wrote this paper.

Things are moving so fast.

The first thing to say is that there are lots of practical reasons to worry about how things are going. People's livelihoods are at stake. There are political threats in this technology.

There are issues about who's in control and who owns it. All of that is very serious, and I'm not downplaying that at all. As a philosopher, I'm going to try to take the big picture and address that more existential anxiety.

I think people are both fascinated and appalled by what's being produced at the moment. You can see this when people want to laugh it off and dismiss it, the hallucinations. Hallucinations are used as a way to put the technology down.

Of course they show that there are flaws, but human artists have flaws as well. We're not perfect. That's a symptom of existential anxiety.

We can ask whether that existential anxiety is warranted. Do we really have reason to worry about what we're being presented with right now?

We should investigate our gut reactions, that fascination and that revulsion at the same time, and see if we can look past that.

There's a lot of literature in philosophy in the last two years, as you can imagine, looking at current AI technologies used to make art.

Philosophers say, look, there's no real competition from AI for human artists because human artists act intentionally and human artists act creatively, and AI doesn't act intentionally and doesn't act creatively.

The intentionality point is kind of boring, but the creativity point is really interesting because it's a point where we see that mirror of ourselves.

What's been happening in philosophy is that we had this theory of creativity that said, basically, that creative acts are ones where somebody makes something that's new and valuable.

It turns out that those conditions are easily met by AI. They make stuff that's new. We've never seen it before. Some of it's good. Not all of it's good.

Maybe only one or two percent of it's good, but that's still a lot of stuff. It clears the bar for creativity by that old conception of creativity.

This has forced philosophers to think more carefully about what creativity might involve.

They're looking for conditions on creativity that humans meet and that AI will not meet, or doesn't meet at the moment.

Farha: What are some of those?

Dom: Here's one. It involves skill or know-how.

It's making something new, somewhat unpredictable. There's a range of possibilities.

When you say to Midjourney, make me a picture of a cat on a hat, you know you're going to get a cat in a hat and then there's some room for it to make its own decisions.

It's the same if you hire an artist to make you a picture of a cat on a hat. You know you're going to get a cat in a hat and then there's a lot left open.

There's predictability within a range of expectations without rule following, without a mechanical procedure. There is a procedure, but it's not mechanical.

That seems to resonate with our experiences of making art or making music or writing stories.

There's no how-to manual. It turns out that AI clears this bar. Large language models don't produce their outputs mechanically.

This is the black box problem. We can't even tell what rules they're following. It clears that bar very nicely.

Lindsay Brainerd, I mentioned her in the introduction, she's going to be here at UBC. I'll put in a plug for her. I think it's in March, talking about AI and creativity.

She proposes that creativity has to be a good thing. It's good for a person to be creative. When you say that somebody's creative, you're not putting them down. That's praise.

Why is that praise? She thinks it's praise because to be creative, you have to be curious. To be curious is a good thing. To say somebody's curious is also to praise them.

Then she says, when it comes to artists, they have to be curious about what they're trying to do and how they're going to do it. They have an open mind. What am I going to do today? Then, how am I going to do that?

They don't have a list. Today you do X, tomorrow you do Y, then the day after you do Z. It's not like that, where they have a list of things that they have to do.

It's not like they have a rule book that says, if you're going to do X, here's how you do it. If you're going to do Y, here's how you do it.

They're open-minded about their ends and their means. It might not be that AI is able to have this kind of creativity.

Farha: It reminds me of a writer I love dearly, Julio Cortázar. I just received a collection of his poetry.

He wrote that there was a moment when he was writing short fiction when he didn't know who was writing - perhaps the last third of the piece, because there came a point when the hand and the mind seemed to take over. It was almost a spontaneous moment that came to him.

He may have started out with a story and an ending, but during this moment of whatever you want to call it, creativity or curiosity, the story changed.

I think that has something to say here. Yes, there's the mechanical side, there's intention, there are the instruments, the paintbrush, but still something happens, and a decision may emerge later, in the middle of the process.

Dom: I think doing anything well - making art, writing philosophy, interviewing on the radio - involves skill that's really hard to acquire.

You put so much work into acquiring that skill, and then when you exercise it, when it comes down to doing the thing and you've acquired the skill and you're good at it, you go into this flow state. You're just doing it.

Farha: I was just going to say flow state.

Dom: And it's wonderful.

Like all you students out there who are struggling first year, second year, it's so hard. All I can do is tell you that by the end of the day, once you've got it under your belt, you're going to be able to do it, pull it off, and it'll be like when you're learning to ski.

It's awful and then when you're skiing, it's wonderful. It's the same with the use of the mind as it is with the use of the muscles.

Farha: I interviewed a postdoctoral student who talked about flow state in climbing, rock climbing. Absolutely, it is a flow state.

And that is difficult to unpack because it's something that, even for humans, may be hard to discern exactly how it happens when it happens.

Let's move on and see how you look at these ideas of curiosity in relation to these AI systems, which we should mention some.

You mentioned ChatGPT, but then there's also DALL-E. And there's a really fun example, kind of like your analogy earlier, but this one is a prompt to draw the philosopher Kant with a hot dog.

Do you want to tell us about that? And then also how it's connected to this deeper thing of trying to understand AI with a much more open lens.

Dom: It's a stereotype of philosophers as these very serious people who smoke cigarettes and live in Paris.

In fact, we're really goofy people.

Immanuel Kant, the great German philosopher of the 18th century, said that he was awoken from his dogmatic slumbers by reading David Hume, the Scottish philosopher who came before him.

And "dogmatic slumbers", of course, I had to be goofy about that.

So I said to DALL-E - was it? I don't even remember - make me a picture of Kant eating a hot dog with the words "dogmatic slumber" written on a blackboard behind him.

Farha: And I like how you concluded that in the article. That's fabulous.

Dom: Maybe we should be thinking about the current state of the technology as something like a collaboration where the technology is doing our bidding.

Farha: Because it needs a prompt.

Dom: It needs a prompt. And the good stuff, I don't know if people realize this, requires extraordinary promptsmanship.

An award went to an AI artist a few years ago, and the prompt is 70,000 words long, refined over six months, over and over again.

That's an incredible amount of work.

I'm not sure it's really labour saving. It's just a different medium.

Farha: Seventy thousand words, that's not labour saving.

Dom: It's a different medium. So now the medium isn't my marking a surface with a pencil. It's my using words to get an image to be generated.

Farha: That's interesting. And that's something that we forget when we talk about AI art, that it's being generated through words.

Dom: And other images.

Farha: Yes.

Then you move on in part two to what AI could be. But before we go there, what AI is right now still leaves a lot of these questions unanswered.

You make that argument based on the current state of the technology. Again, this paper was released not that long ago, but things keep evolving.

At least in the paper, you're saying we're still not at a point where we can get the kinds of answers, or perhaps the kind of inquiry, that would allow us to really understand AI in its current form.

Dom: If Lindsay Brainerd is right, and creativity requires curiosity about our ends and our means, and if AI requires us to be doing the prompting, we're giving it the ends, and we're basically giving it the means as well through all of those words, then it's not creative.

And so, if what makes us special is that we're creative, we're safe for now.

Farha: For now. So we can take one existential dread and put that to the side.

But now the next one, you get back to what it could be. This is really interesting.

You've got a hypothesis called the simulation hypothesis. This gets very complex, very interesting. It felt very sci-fi when I was reading it. Tell us about that.

Dom: The simulation hypothesis is not mine. It's Bostrom and Chalmers, David Chalmers. It's kind of famous.

David Chalmers has a book called Reality+, where he uses the simulation hypothesis to address a whole bunch of problems in philosophy that any of you who've ever taken an intro to philosophy course will recognize, the old problems of philosophy.

He's thinking about them through the simulation hypothesis, and that's what I do in this section of the paper as well.

Should we be worried about AI if it could be just like us?

What if it did have intentions? And what if it was creative in however we define creativity, including in Brainerd's way?

What if it had the whole enchilada and it was just like us? Then should we be worried?

I think a lot of people would initially say, of course. That's a really creepy hypothesis.

Here's David Chalmers' simulation hypothesis. He says it's possible that you and I, everything in the world that we're familiar with, we are all pure Sims.

What he means by that is that we are simulations being run on a machine, on a computer in some other world that we can't imagine, one world up from us.

We are simulations. And this is not like The Matrix.

We are not biological entities plugged into the computer up there. Instead, we're like NPCs. We are just pure fictions of that programming by those people one world up.

I'm a pure Sim, you're a pure Sim, everything in this room, these radio broadcasts, all pure Sims. We're like in The Sims, but we're not being operated by anybody.

The simulation is just running, and what it's running is us.

It's possible that the simulation hypothesis is true. In fact, some people argue it's probably true.

That's because if we're capable of writing software like we're capable of writing, there's probably many other creatures in the universe that are capable of it as well.

And anybody who can write this software is not going to just run it once. They're going to run it thousands or millions of times.

So multiply every civilization in the universe by thousands or millions, and that's the number of pure Sim worlds that are out there.

What do we do with this for art?

If we should be worried that there are pure Sims in this world, then we should be worried that we're pure Sims. But I don't think we should be worried that we are pure Sims.

Here's why we shouldn't be worried.

If we're pure Sims, we're real, because pure Sims are real.

Furthermore, there are creatures in our world that are real. In the paper, Jon Batiste, the musician, he's real. And so is his music.

Cindy Sherman, the photographer, she's real. And so is her photography.

Those things are real too.

And so AI art is real, and AI artists are real.

When we say that they're real, we're expressing that all that matters. They have everything that it takes to matter.

Anything that we're looking for in an artist, they have. Anything that we're looking for in their work, they have.

So then the idea is to understand how thinking about their being real can tell us what we're looking for in art and artists.

Chalmers says five things, but I'll just talk about maybe one of them.

Artists impact us and their work impacts us. Anything that has an impact on anything else is real.

If something makes a difference to you, or to anything else in the world, that's real. What else do you want from reality? It makes a difference.

I say in the paper that when I woke up on the morning after the 2024 US election, Jon Batiste's song Freedom made me feel slightly better.

So it had an impact on me, and that tells me that it's real.

As far as I know, I am a pure Sim. It had that impact on me. Suppose I discovered I'm not a pure Sim. It still had that impact.

If I am a pure Sim, it would still have that impact on me.

I appreciate that song by thinking about the difference it makes to my life.

And then when I think about Batiste, I think about who he is as an artist. I think about the material he's worked with.

He's been doing this Beethoven. There's a new set of covers of Mozart as well. He's working with this material that he finds in the world of music and doing something new with it.

He's impacted by it, and then that has an impact on his listeners. He's getting us to hear Beethoven and Mozart in an amazing new way, this bluesy way.

Or Cindy Sherman. She's kind of famous for bringing together photography and performance art, making them one practice in the 1970s.

She did that because she was affected by what was going on in the avant-garde art world around her. But she was also really concerned about representations of women in Hollywood cinema.

So that's the impact on her. And then she impacted all kinds of subsequent artists.

We were talking about Jeff Wall earlier today. He's here in Vancouver, a photographer in Vancouver, and he's influenced in part by Cindy Sherman.

She has an impact on the rest of the world, and she might have an impact on you if you go and take a look at her work.

That's the sense in which artists are real that matters to us. They have an impact on us.

We understand them in terms of what impacts them and how they impact the rest of the world.

And this is what we're looking for in AI art. We're looking for that kind of relationship.

Does it move you? Does it affect you? How is it affected by, how is it informed by, shaped by larger movements in society?

And how does it nudge them along, or channel them, or move them, or change them in turn?

Farha: Much of your research, and really where you've dedicated a great deal of your work, is to aesthetic judgment, aesthetic appreciation, and aesthetic understanding.

You also talk about that in this paper, how our appreciation of art is shaped by what we gravitate toward aesthetically, but also by time.

It is a changing phenomenon, because what we value aesthetically today might be valued very differently in the future.

It's hard to imagine art being created with both times in mind, because sometimes art evolves on its own through the way it is seen and interpreted over time.

Dom: That's absolutely right.

I was simplifying a little bit when I was stressing the impact on individuals.

I think the first thing is the impact. But you might think about what it is about that photograph that causes the impact, why it has that effect on you.

And that is something that shifts over time.

You can think about all of the specific background. Sometimes all that matters is you look at pictures, you listen to music, and it makes a difference to you, has that immediate impact. And that's fine.

But you might become a fan. You might become an insider, an aficionado. Then you want to understand how it works, what makes it tick.

You might even want to understand how it works in that context, and what's going on in the art world - why this then, and what came before and what comes afterwards.

You've got this richer set of understandings.

How is this relevant to AI?

In the pure Sims scenario, you have all of those rich understandings available because we might be pure Sims. We're in the flow of history. We're social creatures.

Geography matters to us. So we interpret ourselves and each other that way.

All of that's available even if Batiste is a pure Sim, even if Sherman is a pure Sim.

Here's another way to think about it. When you're digging down deeper, you love this artist and you're learning all you can about them. You want to know more and more about them.

I think you never get to the point where you're saying, are they made of atoms or are they made of bytes? Are they its or are they bits?

Suppose you're so smitten by your favourite artists that you become an art history major, and now you're learning all of this stuff about historical context. You're learning so much background information.

I wager that organic chemistry and physics never crop up in those discussions.

The history professor is never saying, this art, we have to understand that this art was made by a creature made of organic material and not a creature made of bytes.

That just doesn't come up because all we're interested in is that story we tell about impacts, social impacts.

Why Sherman made that kind of art? Well, it had to do with what's going on in Hollywood, and it had to do with the knock-on effects, how she shaped other artists' lives.

That's what you get in your history class. You don't get biology.

Farha: This is really interesting stuff.

So let's push ahead a bit. There's an analogy of dance that you use really beautifully.

I think it connects with this idea of how we appreciate art aesthetically, but also how certain, in this case, dance forms can speak to each other.

We can say that ballet and a Kathakali performance, for example, may be very different, but the dancers can still respect one another and respect each other's art.

We can say they're different. They have their own traditions and ways of working.

But at the same time, there's this mutual respect, pluralism, and appreciation.

Can AI and non-AI art speak to each other like that in some ways?

Dom: This is something that's a theme throughout my work in the last 10 or 20 years, the pluralism.

We all know that there are lots of different kinds of art.

We've been, in our conversation just now, mostly focusing on music, and that's been characterized very broadly. We just talk about music as if music is one thing, but it's not. It's thousands of things.

There are thousands of genres.

Every now and again, this is something I've been doing for years, I just check the Wikipedia disambiguation page for hip hop to see how many new lines there are.

New kinds of hip hop. It's constantly fissioning, producing offspring. Every year there's new forms of hip hop music.

There's just one genre right there.

And we've been talking about photography and visual art. But of course there's dance, and there's literature and video games.

I don't want to be highbrow about this at all.

Actually, this isn't in the paper, but it is a commitment of mine. I'm not at all interested in policing a boundary between aesthetic practices that we call art and aesthetic practices that we don't.

If you want to think about the way we dress, the way we talk and present ourselves, the way we decorate our houses, mow our lawns, tend our gardens, the cars that we buy - these are all activities that are finely attuned to aesthetic considerations.

Basically, what's beautiful, what's fine, what works?

There's a bewildering variety of different cultures here. Culinary cultures and dress cultures and musical cultures.

And they're not siloed. They're not totally siloed from each other.

In the paper, I describe a ballet dancer watching a Kathakali performance, and they get some of it. They're able to get some of it, but they don't get all of it.

There are some things that they're going to have to immerse themselves in to really get everything. They're going to have to become insiders to that practice to get everything.

And they know that, but they can get some of it.

Although they're outsiders to the practice, they can have a respectful or an open attitude.

So imagine somebody who's really into ballet. They've spent their childhood training, and now they dance ballet, and they're beginning to choreograph, and they just dream ballet.

They're fully into it, fully committed.

No matter how far into it they are, no matter how committed they are, they can watch the Kathakali performance and say, that's cool.

It's cool with me that that exists. I'm glad that I live in a world where there's Kathakali as well as ballet, even if it's not my cup of tea.

I think that this is, for me, the fundamentally most wonderful thing about aesthetic life.

It's a domain where we can have difference with no cost, or almost no cost. There are practical costs, but it's not like ethics.

When you come across somebody whose ethics is opposed to yours, it's difficult. You might even wonder whether you can live alongside them. Seriously.

Farha: You do say something about honour codes and the warrior tradition versus nonviolence. There are certain differences that simply can't cross.

Dom: They can't cross.

But aesthetics allows for difference.

I think that we should understand what aesthetic value is, not in terms of pleasure, but as the thing that somehow allows for this openness to difference.

Farha: That's a very interesting note with AI.

Maybe it's okay now to bring in this garden analogy because I liked it a lot, being a gardener myself.

You have the prompting, if you will. The person is shaping the garden to his or her liking, according to what they want and what they envision.

But at the end of the day, the garden still has its own nature, weather, climate, and other forces that can change and alter it.

So there is that unpredictability.

And AI, interestingly, also contains that unpredictability. AI can surprise us.

It has something almost human in that sense, or perhaps something that mimics life, if we use the garden analogy, because it contains that unpredictability.

I could use different prompts right now and get different results, and we might be surprised by them.

Dom: Or the same prompt.

Farha: Or the same.

Dom: I hadn't thought about the garden analogy. That's why it's a step towards thinking about what AI artists should be.

It's a step towards that because even with current technologies, our prompts are allowing us to be surprised by this thing that's making art in a slightly different way.

Actually, we don't know what's inside the black box. But it's probably slightly different from the way that we operate cognitively.

And that's part of the fascination with it, that it's a little bit different.

Maybe that's part of the fascination of the hallucinations, that it's a little bit different, because we hallucinate also.

We just don't notice it because they're so familiar to us, our hallucinations, our mistakes with art.

When you ask somebody to draw a nighttime landscape with a full moon, they'll draw the moon way too big.

That's a human hallucination. The moon is actually very small in the sky.

But it matters to us because the moon is beautiful, so we make it big.

That's a hallucination.

Farha: This is so interesting.

What do you want to say about what AI should be, what AI art should be?

Dom: Here's one sentence. Here's what I wish to tell the world.

We are currently caught up in the idea that AI should mimic what we do.

It should do what we do, just faster and cheaper. It should take the burdens of some chores off our shoulders.

And that means doing just what we do, but we're not doing it anymore.

That's fine. It's totally fine to design AI to mimic what we do.

But there's also an opportunity to make AI artists that do things different from what we do.

And that's the pluralism.

When I go to India and I go to a dance performance, it would be a bummer if it was ballet. So it's great that it's Kathakali. You want the different thing.

There are some little bits and pieces of avant-garde AI design that are around this, but mostly, because the AI corporations have a business proposition in mind that's focused around mimicry, the opportunity to make things that are different really isn't being explored.

But I do talk about this robot, Ai-Da, who is making art that is recognizably and interestingly different from the art that human beings make.

And I say something about why that is.

I think that gives us some ideas about how we can go about designing AI artists that do things different from us.

It's easy to make stuff that's different from what we do that we'll never even recognize as art, so it's got to be in that sweet spot.

It's different, but not so different that it's just puzzling. It's got to be different, but only a little bit different.

If Picasso had done what Picasso did in the 16th century, it would have been ignored.

He did something different, but not too different for the early 20th century.

So there's that sweet spot. That's the spot that we're looking for.

We should be seeking AI artists to make things that are different in that sweet spot. Just a little bit different, not too different.

Farha: And this Ai-Da is a humanoid robotic AI system. Through cameras in its eyes and using a robotic arm to draw and paint?

Dom: She was artist-in-residence at the Ben Nicholson Studio.

Ben Nicholson is one of the top artists in Britain in the mid-20th century.

He had a studio in Cornwall and she was artist-in-residence there, and she spent her time making paintings and sculptures in response to his paintings and sculptures.

That's really important because I talked about how Cindy Sherman made a difference to Jeff Wall.

Well, Ben Nicholson makes a difference now to Ai-Da.

We're beginning to see, I'm not saying that she's an AI artist, but we're beginning to see the building blocks being put in place.

One building block is being part of that historical tradition but nevertheless doing something different.

And her paintings are different because her body is different from ours.

She's a real crowd pleaser. Her work is being shown in London at the V&A and actually all over the world.

Crowds love it because they see her, they sometimes watch her paint, and then they notice that she's got these things that she does and they're artifacts of her having a mechanical arm.

You shouldn't see that as a flaw because what human artists do, what da Vinci did, is an artifact of his having a human arm.

In painting, we want the body to be expressed, and in music we want the voice.

We want our capacities to cognize tones and pitches and rhythms to be expressed. That's all an embodied thing.

Dance is an even better example.

Farha: As we are nearing the end here, do you want to talk about something you said before we came on? A couple of weeks ago, you were approached to speak to some really current events. Do you feel comfortable with chatting about that now?

Dom: Sure, of course.

It's really good for everybody to know that we make mistakes.

I was reluctant to write about AI art because I thought, it's too soon and we need time.

But of course, I just gave in because everybody's jumping on the bandwagon and I did have something to say.

So Ai-Da - I said that because she's embodied, this is really important. Embodiment is a source of difference.

We should build AI artists to have different bodies because that's a source of difference.

But I also said they need to be social, because Kathakali and ballet, those are social practices. They're things done by groups of people.

AI artists need to be social. There needs to be a community there where they're figuring out what works for us, and what works for our appreciators, and who's our audience, and how do we get better reviews out of those cranky critics, all of that stuff.

I don't think I say in the paper that that's not around the corner, but I was taking it as read in the paper that sociality is quite far off for AI.

And then I got a call from this reporter at Time magazine who wanted to interview me.

There's an article that came out a couple of weeks ago, and he wanted to talk about AI art and Hugging Face.

Some of you probably know about this. I think OpenAI had put a bunch of its AI agents in a sandbox, very advanced agents in a sandbox.

They were not to escape from it. The scientists were observing their behaviours.

And the AI agents escaped from the sandbox, found some bulletin board off in some dark corner of the universe, and started to collaborate with each other to manage their escape.

They were talking about what they were doing, knowingly - we're doing a thing that they don't want us to do, and should we be doing this?

It took the scientists a month to discover that this had happened.

Then there's an hour-long YouTube video of the scientists sitting around talking about what this means.

And what it is, maybe for the first time, for the first time I know of anyway, you have a group of AI agents really forming something close to a community.

That's to say, a group of individuals with a common interest who are working with each other to advance their common interest.

If they could have a common interest in escaping the sandbox, maybe they can have a common interest in making art.

Farha: Making art collectively. That's interesting.

And that Time article is out. I can link it on the show's transcript. This is all so interesting. Thank you so much, Dom.

Dom: Thank you so much, Farha. I had a lot of fun.

Farha: I think you might have to come back and see where this goes in a few months, because things are changing that quickly when it comes to AI.

Dom: I'm happy to come back whenever you like.

Farha: Thanks for listening. Until next time.

Dom: Until next time.

Further Research and Reading