CohentheWriter chats with the Digital Art Pioneer
The following is the condensed and lightly-edited transcription of a conversation between Anne Spalter and Max Cohen (@cohenthewriter) recorded in September 2022. This is part 2 of 3. Read part 1 here.
Max: When were you exposed to generative art for the first time?
Anne: When I was writing my book. I wasn’t creating generative art at the time because I’m a slob, which makes me a very bad programmer. So I was studying programming, and I was communicating with programmers, but I was not great at writing my own software. But working on the book and talking to the pioneers in the field like Frieder Nake and Roman Verostko and Manfred Mohr, and seeing their work, it really grew on me, and I began to appreciate the beauty of it. That said, I didn’t see what we now call “Long-form Generative Artwork” until the blockchain…and Artblocks!
The earliest generative artists would write an algorithm that would produce many outputs, and then select the ones that they thought really represented what they wanted to do, and they wouldn’t worry about the many outputs that weren’t above threshold aesthetically. The algorithms didn’t have to produce 1000 outputs that were all great looking, like so many today, they only had to produce five or 10 or 20 or something.
Max: My larger question is, “What is your process?” but I suppose I mean more like, philosophically, how do your ideas become distilled into these art pieces of yours, and what do you use to turn them into legitimate pieces of existing art?
Anne: I work with a lot of different media, so the process is sometimes different depending on the medium. But I would say the content of the work has been the same in terms of the imagery and a lot of the meaning of the works. Since I started, really in high school and when I was an undergraduate, the imagery that’s in the work, the highways and lighthouses and outer space stuff have been incredibly consistent throughout my whole career. And I think it’s all — okay, so I’m really into Carl Jung. My hope is that, even though they’re personal symbols, they’re also things that resonate with the collective unconscious by having meaning across different cultures and times. Like there’s always something particularly modern in my work that has some counterpart throughout history, some sort of signaling device — like a lighthouse, which actually has existed throughout time, or even UFOs or things in the sky have been drawn throughout many time periods.
So I think the imagery is very consistent and the exact process depends a lot on the context that the piece is made for. So if it’s a public art commission or a show, I might be responding to a specific place or a theme, or a medium if it’s online, and that would help determine where I would start.
I usually start digitally though, with photographs or things I’ve collected. I take all my own photos and videos, so often it starts with that. Recently I’ve been using AI, which has been super fun. And very futuristic. People who have trouble conceptualizing relatively simple algorithms, I think their minds are just blown by the whole concept of AI. When someone first told me you could type in a sentence and get a picture, I didn’t believe it. And now I’m using that every day, different Txt2img websites, Discords, all the different things that are out there. And it’s just so bizarre. But incredibly fun and addictive. Sometimes I’m up at three in the morning thinking, I’m just going to do one more, and then you see it sort of come to life like a Polaroid. And you get some kind of wonderful image. And then you’re like, “Oh I’m going to tweak my prompt a little bit,” and hours and hours go by.
Max: Recently, the actual workings of how AI software works, the Txt2img pairs and prompt engineering, was all revealed to me. As a person who, at the beginning, was admittedly a little bit dubious about AI artwork in general… I do want to talk about your work specifically.
I went to your website, and I was looking through all this artwork, and of course it’s all these edifices you mentioned: planes, ships, bridges —
Anne: Modern landscapes. Transportation. Weird modern stuff you see in the landscape.
Max: It’s a little menacing to me, if that makes sense.
Anne: Well, it has no people. And it’s kind of always been like that. Although the show I just installed does have people.
Max: What draws you to human edifices?
Anne: I don’t really know. It’s hard sometimes to explain the work, but it’s definitely the work I have to make. I know that because it’s what I always make. So I think it’s sort of an exploration, like what’s out there? What’s beyond? Exploring.
Max: I used to have dreams about bridges when I was a kid, bridge nightmares, especially the Verrazano bridge. Even still, when I go over it my heart drops into my chest.
Anne: I’m obsessed with the Whitestone Bridge. I take photos of it every time I cross it.
Max: Why is that?
Anne: I don’t know. I just really like it. It’s like you’re getting somewhere… else.
Max: There is something that seems inherently oxymoronic about enlisting an AI to create impressions of items which are quintessentially human. I was wondering if you had any thoughts about that dichotomy.
Anne: So, as I mentioned, I’m very into Carl Jung. And he says, of course, people have this sort of collective unconscious. And I love the whole Txt2img thing because it’s based on millions of images, so it is all kind of a cultural collective unconscious. AI art is kind of tapping-into all these images, and from the internet mostly. So it’s like a weird slice of cultural collective unconscious from, I don’t know, 2005–2012, depending on which image database your program is using. I’m fascinated by that.
Max: What are your preferences of program?
Anne: Different ones are good for different things. I’ve been using everything from StableDiffusion to Dall-E to Midjourney, and ColabNotebooks. Everything. I love Midjourney for its artistic output. You can really make things that look like paintings, and it sometimes gives very unexpected and beautiful compositions. I think you can just put words together, and especially if you’re making kind of serialistic things, it’ll combine them in beautiful ways.
I found that with Dall-E and StableDiffusion, you have to be a lot more specific about exactly what you want in the composition, and if you just string some words together sometimes it doesn’t know how to handle it. So you kind of have to talk to them in different ways. Each one has a learning curve in terms of the prompts, so I like all of them, but they’re sort of different.
Max: Do you think it’s an overall net-positive or an overall net-negative that things like Dall-E and Midjourney and the Discord-esque AI software are just kind of there for the layperson to start experimenting with?
Anne: I mean, I think it’s great. Especially as an educator. I think a lot of people feel like they can’t make art: They have to learn how to draw, and drawing is super hard, and they might never learn how to do it, and it’s all very off-putting. So I think it’s wonderful for people to be able to make really compelling images easily and thus communicate visually. I think that’s a net-positive. And I don’t think there should be a lot of gatekeeping in visual creation and communication.
Max: Very crypto art: Get rid of as much gatekeeping as possible.
Anne: Yeah! And I’ve even sometimes made images with AI just to — almost like the way people use emojis, like I want to say something and I’ll just type it in as a prompt and send it as an image. So maybe people will do that in the future.
Max: I’ve never thought of that before, it’s fascinating. Like Txt2img software within the text browser itself.
So one thing that stuck out to me about your work is that so many of the images you create, your AI Spaceships for example, they’re juuuuuuust recognizable enough to be recognizable. It seems like you achieve a pretty fine balance between the large amount of abstraction you allow into each of these images and an observer’s ability to come in and pick out a recognizable image within. Is that something you intend to do? Or is it just kind of an effect of how you use the software?
Anne: I feel like no matter what software I’m using, all my work sort of looks the same. Which is funny because there’s a lot of discussion — a lot of long Twitter threads — about, “Is AI artwork artwork?” and, “What about people who just come in and get lucky and write a prompt and sell a piece of artwork and someone buys it, but they’re not really an artist?”
I’ve come in on the side of: AI is a perfectly legitimate art tool, and part of my reasoning is because I feel like I’ve just continued to make all my same artwork with it. It looks like my artwork to me, it has the same style even though the process couldn’t be more different from really traditional drawing. If I make a charcoal drawing or a painting, for instance, it’s the same tools we used in the Renaissance. But now here’s this tool that is totally particular to our time, and yet the imagery is the same. So it’s something that comes from your brain, your heart, somehow, as an artist. It all comes from you as an artist.