[music]

Alison Stewart: This is All Of It on WNYC, I'm Alison Stewart. A new exhibit tells the story of one historic convention in the world of art-making, artificial intelligence. Today you can go online and find all kinds of AI programs that can generate art based on user input, but back in the 1960s, years before the AI world we know today, artist Harold Cohen dreamed up AARON, an AI software designed to make art, the first of its kind. Cohen conceived of the idea while working at UC San Diego. The story of his creation is a subject of a new exhibit at the Whitney. In the show, you'll see original drawings created by AARON, Cohen's own sketchbooks detailing his process of developing the program, and even live drawing demonstrations in the gallery featuring modernized recreations of the early stages of AARON.

Joining me now to discuss the show, which is on view through May 19th, is Christiane Paul, Whitney curator of digital art. Welcome to the studio.

Christiane Paul: Thanks so much for having me.

Alison Stewart: I think it's so interesting that Harold Cohen was an artist, not a tech person. [chuckles] He was a fairly well-known and appreciated artist, he'd been to the Venice Biennale. How would you describe his work? How was he viewed as an artist before coming to the States? He's from England, I should say.

Christiane Paul: He was a very established artist. As you mentioned, he represented the UK in the 1966 Venice Biennale. He had shows at the Tate and at various institutions and was really known as an abstract painter, but he also became a little bit frustrated with painting. When he comes to UCSD in the late 1960s, 1968, he gets introduced to programming.

Alison Stewart: What was the general landscape of programming and artificial intelligence at the time and its relationship with creativity? Were people thinking about the two together?

Christiane Paul: Yes, they did but Harold's work still is very groundbreaking. Many people ask me, "Oh, did AI even exist at the time?" As a field, it was formalized in the '50s, John McCarthy coins the term, and there's the famous Dartmouth workshop. It was an early field at the time, but I think Cohen was really pioneering in exploring it in relationship to creativity.

Alison Stewart: Where did he learn to create technology like this?

Christiane Paul: He was really absolutely self-taught. He taught himself everything from scratch, so the AI he creates is early symbolic AI, which was the AI of the time, and that is really rule-based AI. His soul system depends on complex rule sets, and he had to write all of these rules and create them from scratch, so very different from today's programs.

Alison Stewart: The idea that today's programs, they mine the internet for information. In this case, it was Harold Cohen's information. He created it, he created the rules. Is that the way? Am I describing it correctly?

Christiane Paul: Absolutely. He gave AARON his software knowledge of the world and the objects in it and internal knowledge about how an artist would represent that world. AARON knows, for example, that a human has a face with a mouth and a nose and eyes and hair in different styles, and that there can be different poses. Harold really had to "teach" the software or program all the rules for proportionally positioning that, all the rules about doing a successful composition with foreground and background. Once again, as you say, very different from today's statistical AI, which has been trained on massive data sets. Without understanding any of these rules of composition is basically just mashing it together to present you with a cactus on Mars in the style of Monet. [laughs]

Alison Stewart: Can you describe an example in the show where we see Cohen's programming of AARON to understand perspective?

Christiane Paul: What I really wanted to foreground in this exhibition is that the software is at the core of it all. The paintings and drawings created by AARON have been collected by institutions worldwide, but the Whitney is the first institution that is collecting the software over time. You can see the process on the walls, we have one version of AARON projected that is figurative, that was actually done in collaboration with Ray Kurzweil, the computer scientist, and once released as a screensaver.

You can see basically AARON at work creating compositions, and that very same software, we are also plotting in black and white on one of the plotters in the adjacent gallery. It is actually for the first time since the '90s that AARON's output is plotted life again. Harold Cohen not only created the software but also built drawing machines and drawing devices, and in the '90s, a painting machine to execute the work.

Alison Stewart: Yes. I want to describe for people, you go through the exhibition, and then at the end, they're the actual machines and you're watching AARON work. People have a lot of questions. I went and people had a ton of questions. Luke Green, our producer, people had a ton of questions. What do people want to know about the machines when they watch AARON at work?

Christiane Paul: Oh, there are so many things. First of all, are these Harold Cohen's drawing machines? They are actually not. We looked into resurrecting the carcasses of those drawing machines but quickly realized that in the process, we would really destroy them, or it would take at least a decade probably to revive them. These are built by Bantam Tools. Bri Pet is one of the members of our digital art acquisition committee, but we tried to stick as close to the original as possible. We looked into the paper sizes, we looked into the line width and how things are drawn. We are really trying to make it as accurate as possible historically.

Another big question is always, or I see people looking at it and saying, "Oh, it's not art, it's the machine doing it." Either the plotter or the software, I'm not sure what they are thinking, but once again, it's important to keep in mind that Harold really wrote that software as opposed to the corporate tools you're using today. If you write a poem and then print it on your printer, it's not the printer that wrote the poem, it's still you.

Alison Stewart: That's such a good description. My guest is Christiane Paul, Whitney's curator of digital art. We're talking about the new exhibition at the Whitney, Harold Cohen: AARON, it is up through May 19th. I kept looking for the acronym like, what does AARON stand for? It's not an acronym. Will you share where AARON comes from?

Christiane Paul: Yes. People ask me that question all the time. There is actually frustratingly not a definitive answer to it, but in a conversation with Pamela McCorduck, who wrote a book on AARON, Harold mentions that there's definitely an allusion to the biblical character of Aaron, Moses' brother, who, according to the Bible, was more eloquent and was anointed at Moses' mouthpiece.

Harold would refer to AARON as his other self, so here is the character speaking for him. Harold also was very critical of the mythification of artists and their divine inspiration. He always wanted to undermine that. What we do not necessarily foreground is that AARON also was Harold's Hebrew name, Aaron Mordecai Cohen is actually his Hebrew name, but that's also not why he named it AARON, because the original plan was to actually give each version of AARON a different name going through the alphabet, going from A to B to C, which, in the end, he did not do.

Alison Stewart: That's so interesting. Oh, that's so interesting. Did he receive pushback initially when he started, when he presented AARON to colleagues, to people in the art world?

Christiane Paul: Yes, absolutely. I mean, it was groundbreaking what he did at the time. He worked within the community at UCSD and received an invitation to the Stanford Artificial Intelligence Lab in 1973, where he also officially named AARON. He was also surrounded by scientists who would've understood creativity in that context. The attitude of the art world at the time towards all of the artists who started using computers could be downright hostile. Of course, people would question what it is. Harold would say, actually, most people would ask, why is it art, and not necessarily say it isn't art? Although I've heard the latter two in the galleries [laughs] over the past few days.

Alison Stewart: Why is it art? I'll play that game. Why is it art? [laughs]

Christiane Paul: To me, it actually is the highest form of art. What Harold once said was, "I felt if I could encode a system, an artificial intelligence system with the knowledge of how to represent and create art, I would learn more about painting and drawing than I ever could just doing a painting or a drawing." I think this way of really encapsulating what artistic representation means into a system to me as meta art. Harold always made art about the meaning of art. I consider it one of the highest forms of art.

Alison Stewart: It's so interesting when you think about that because to create the programming, he had really had to think about what art is. He had to be able to articulate it so that AARON could then understand what art is and then make it.

Christiane Paul: Absolutely. To one, what was at the core of it for Harold was intention and intentionality. For him, art was about the intentionality of the line. The earliest form stage of Aaron he developed was the freehand line algorithm. Everything looks like it is drawn by hand. It's not the typical computer-generated line of A to B to C straight line. Everything looks like it has been hand-drawn. That is really at the core of AARON.

Alison Stewart: He also, I understand that Harold Cohen wanted the program to have a childlike sense, like a sense of early stages of when children start to draw, why would he want AARON to mimic the drawing process of a child?

Christiane Paul: Yes. One of the essential questions that Harold asked himself was, when does a mark generate meaning? Where is that transition being made? He was very interested in how children learn to draw and how they do the weird squiggles, but then suddenly a line is emerging and suddenly the line turns into form.

He built the early stages, the more abstract stages of AARON around that. He also was very inspired by a visit in 1973 to Chavon Valley in California where he saw the petroglyphs, also these abstract childlike drawings. Those really made their way into the early stages of AARON, the abstract versions before the software became really figurative

Alison Stewart: We're discussing Harold Cohen: AARON, which is on view at the Whitney through May 19th. My guest is Christiane Paul, Whitney curator of digital art. At the exhibition, you can see some of the art, you see the projections, you see the machines in the room doing the line drawings, but then we also see his sketchbooks and sketches the development of AARON. What can we learn or what have you learned about Harold Cohen's process by looking at his sketchbooks, looking at his initial early thoughts about AARON?

Christiane Paul: It was really amazing to have the privilege to go through them. I spent time at the studio in Encinitas going through all of these notebooks. Of course, we can all exhibit a few pages of them, but what you will see in them is, for example, a breakdown of all of the poses that the program will parse through in order to position or create a person in a specific type of pose.

What the notebooks also nicely illustrate is what he was really up against at the time in the '70s, like doing something on a computer, or later on, on screens. The memory was so low that you just ultimately had a certain amount of pixels. Working with that limited amount of pixels and figuring out how they could be layered to even create a form, on paper or on a screen, was really mind-blowing. Today, we do not even think about it, and people say like, "Was that even possible in the '70s?" What went into that is really amazing. We see that in the notebooks.

Alison Stewart: We learned that 1989 was an important year for AARON's development. What happened in '89? How did it help AARON evolve?

Christiane Paul: Around that year, AARON became really fully figurative, and the figures also are three-dimensional in space. We have a couple of the earlier paintings in the show, and you see that the figures are still very abstract. He was inspired by Matisse at the time. At that point, he really moves into figurative representation. He was also really interested in making the work available to the public.

He would give things away. In the early '90s, he starts developing color within the system. He works on a silicon graphics station to have AARON do color. He mentions in one of the exhibition brochures that there was a short phase when he would use those color compositions that AARON would do onscreen to actually paint them and translate them into paintings.

Alison Stewart: Oh, wow.

Christiane Paul: Yes. Those were, I think, only a few years. Starting in 1995, he used an actual painting machine, so that one would not only make decisions about colors but actually apply them from little buckets onto the paper.

Alison Stewart: That would be wild to watch. As curator of digital art at the Whitney, what would you say to someone who's skeptical, who doesn't think that AI and art that they should be in the same sentence?

Christiane Paul: Yes, I get that all the time, and I completely understand the questions surrounding creativity that AI raises. I think most people are looking at the current AI tools, and it's very questionable what comes out as art if the average person just puts in a prompt. I would also say that artists are doing amazing work with AI tools, but one shouldn't underestimate what work they're putting into it. Training the AIs on data sets, endless tweaking to filter out the more questionable results you are getting here.

Alison Stewart: Right. The name of the exhibition is Harold Cohen: AARON. It's at the Whitney through May 19th. Christiane Paul is the Whitney curator of digital art. Christiane, thank you for coming in.

Christiane Paul: Thanks so much. Keep in mind, Whitney is free on Fridays.

Alison Stewart: That's right.

Christiane Paul: On second Sundays. Upcoming weekend.

Alison Stewart: That's right. That's All Of It. I'm Alison Stewart, I appreciate you listening and I appreciate you. I'll meet you back here sometime tomorrow.

[music]

 

Copyright © 2024 New York Public Radio. All rights reserved. Visit our website terms of use at www.wnyc.org for further information.

New York Public Radio transcripts are created on a rush deadline, often by contractors. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of New York Public Radio’s programming is the audio record.