Alan Turing once said, “We may hope that machines will eventually compete with men in all purely intellectual fields.”
We may not be there yet — machines still cannot create content autonomously or intentionally — but artists across disciplines and media are finding ways to implement new technology in their work. Allison Parrish, for example, writes poetry via programmed algorithms; Annie Dorsen uses machine learning to create theater productions. As artists harness the powers of technology for their art, several essential questions arise. What does it mean to create art with artificial intelligence? Are these techniques truly new? And why do we even need art that uses algorithms? This seven-minute episode will explore these questions, among others.
Music: Robot Waltz by Ketsa, licensed.
Jonathan Moens: I forced a bot to watch over 1,000 hours of Seinfeld episodes and then asked it to write an episode of its own. Here is the first page.
Automated Bot Voice: George and Elaine sit on the couch. Jerry eats a plate of cereal. George…
Hannah Seo: You may be familiar with this viral Twitter meme format. The idea is that a bot consumes content, learns from it, and spews out a script for similar content, which ends up being a touch off the mark and yet wildly funny. This meme format began as early as two whole years ago, and while it is definitely fake — no bots actually produced any scripts just from watching videos — the meme alludes to a very real and growing phenomenon.
Allison Parrish: I think these techniques actually have a really, really long history.
Hannah Seo: That’s Allison Parrish. She’s an assistant arts professor at NYU’s Interactive Media Arts program.
Allison Parrish: My practice is mainly focused on text. What I do is make computer-generated poetry, it’s been the main focus of my arts practice for the past 10 years or so.
Hannah Seo: Allison has published a book of computer-generated poetry. She fed a statistical modeling program lines of poetry from a big database. Then the program scored each line based on its phonetics and sequenced them based on how close the lines sounded to each other.
Allison Parrish: I don’t actually believe that these techniques are new. If you look at the history of like, specifically computer-generated poetry, like, it goes back, you know, 70 years.
If you think about language games, like you know, Pig Latin, or you think about games like Scrabble or things like that, letter blocks are a kind of algorithmic text composition. And I really just view myself as being one chain in that long link of poets and creative writers that have been working with chance, randomness and things like that.
Hannah Seo: Allison says that methods of randomness — like drawing words out of a hat to make a poem — were used in the 20th century and prior. Her algorithm, she says, is doing something pretty similar, but the computer is doing the selecting, and at a faster rate. An algorithm, she says, is just a protocol, a methodology.
And it’s a methodology that can be applied to all sorts of art forms, including theater.
Annie Dorsen: The first piece I made like this had no human performers at all, it was a piece for two computers. It was a chatbot performance, and the text was spoken by a text-to-speech synthesizer.
Hannah Seo: That’s Annie Dorsen.
Annie Dorsen: I’m a theater director who works with computer programmers to write and design algorithms that then perform in algorithmic theater.
Hannah Seo: By “algorithmic theater”, Annie means using algorithms to generate new scripts for her performances. She’s created a number of theater productions over the years that run this way. Each show is completely unique, with the algorithm generating new scripts or music as the show goes on.
Bot 1: Can machines think?
Bot 2: I love games.
Bot 1: You are going to lose.
Bot 2: Thanks a lot for telling me I am going to lose.
Bot 1: You’re very welcome.
Bot 2: I think we better define our terms.
Hannah Seo: That was the chatbot piece that Annie was talking about, called “Hello Hi There”. It consisted of two chatbots whose lines were generated from algorithms that Annie programmed. The algorithm was designed to scrape text from a famous 1971 debate between Noam Chomsky and Michel Foucault, as well as comments from YouTube videos for the debate and excerpts from other philosophical texts. She says the algorithm she used was fairly simple for a reason…
Annie Dorsen: I wanted the audience to maybe be able to reverse engineer in their minds how the algorithm is resequencing the text and what the linguistic rules might be. So that was part of that goal of having the audience be able to sort of understand or try to figure it out, to empower audiences really.
Hannah Seo: With artificial intelligence, a computer is pretending to be a human. In theater, a human is pretending to be a different human. Because of that parallel, Annie knew AI (artificial intelligence) and theater were a perfect pair, especially in today’s oddly dystopian world.
Annie Dorsen: I think one issue we’re having, culturally, is that these tools are so complex, they’re so mysterious to so many people, that we feel a little helpless in the face of them. So I was always trying to — I am always trying to create pieces that really engage the imagination of the audience, which means they have to be able to think alongside the piece.
Hannah Seo: In a traditional theater show, you watch actors perform memorized lines that never change between performances. But in one of Annie’s shows, you see a performance evolve according to how the algorithm runs. A lot of what Annie does is bridging the gap between her audience and the weird, complicated world of programming. She likes to quote Arthur C. Clarke.
Annie Dorsen: Any sufficiently advanced technology is indistinguishable from magic, and I always wanted to avoid that magic feeling. I sometimes refer to art that’s about what the technology can do as kind of trade show art. For those kinds of works, you can only have one response, which is “wow, that’s so cool.”
Hannah Seo: Allison, however, has a different take.
Allison Parrish: A lot of my work has also been accused of being a tech demo. But, I’m fine with that. If you’re disparaging something for being a tech demo, then you’re disparaging the skill that goes into producing new forms of technology. And you’re also saying that that’s not a valid way of being expressive, which I don’t think is the case.
Hannah Seo: Allison says that a lot of people have a hard time viewing programming and algorithms as valid tools for creativity and expression. They often think that, because computers are machines, they must be objective, but she says that’s not true.
Allison Parrish: The idea that computers themselves are agents is a lie. And that’s why I’m very, very careful in my own work when I publish a book of poetry, it has Allison Parrish on the cover. It doesn’t say, like, Allison Parrish and poetry-bot 3000 because it’s my poetry. I wrote it, it was my intention. And I should get the credit if it’s good, and I should get the blame if it’s bad.
Hannah Seo: The use of algorithms and programming in art and creativity is an interesting, thought-provoking, and evolving area of work. But why do we even need AI in art?
Annie Dorsen: Technology is affecting our relationships. It’s affecting our economies, it’s affecting our choices, it’s affecting how we know what we know. Why do you need a novel? Why do you need a film? You need it because you’re curious about the world.
Hannah Seo: Making art that uses and comments on algorithms and AI is important because algorithms and AI in life are important. And perhaps with this art, we can better understand the technological world around us, and what it means to be a human living in it.
For Scienceline, I’m Hannah Seo.