In An Uncanny Valley, Art Evolves

AARON ROWE: For the rest of the hour, a quick trip into the uncanny valley and some computer generated, or rather, computer evolved art. It’s up on our website in our sci-arts video. Here to tell us about evolved artists, SciFri’s video producer, Luke Groskin. Hi, Luke.

LUKE GROSKIN: Hi, Aaron.

AARON ROWE: What is this? How does somebody evolve art? What does that mean exactly?

LUKE GROSKIN: Yeah, that’s kind of the question, right? It’s not like you can send a photo or sculpture out into the wilderness and wait a decade and see what becomes of its offspring. It doesn’t really work that way, but what you can do is you can evolve art digitally using computers, and that’s what artist Joel Simon did. He’s an artist, a programmer. He made this website called Artbreeder, and the ideas behind it are really ingenious. He’s taken biology and plugged it into a digital interface.

So what he did was he built a system where users can take a photo or an artwork of portrait or a landscape or something like that, and you blend it with another artwork of the same type, and when you blend it, the offspring is the average between the two, and that means not just visually but, like, on a data level. And when you do that, it’s pretty much like breeding, and now you can do that in three different ways.

You can breed an artwork with itself, randomly selecting the visual genes you want to, that get enhanced, and that’s kind of like asexual selection. One creature breeds with itself to produce more, and then you can breed an artwork with another, and then you get a perfect mix between the two. That’s kind of like sexual selection, and then lastly, you can pick which visual genes you want to enhance in the next generation, and that’s kind of like CRISPR. And when you do this over dozens and dozens of generations, obviously with users guiding the aesthetics, you get basically evolving art. You get, you can see the progress of how art changes over many, many generations. It’s pretty fascinating.

AARON ROWE: And how real do these things look? Are they real, or are they uncanny valley real?

LUKE GROSKIN: Oh, they’re very, I mean the photo ones are super surreal. Like, they have that, you’ve probably seen it online where like people are these semi-computer generated blended mixes. You have, like, if you take a dog mixed with a flower, but then there’s other types of artwork that you can see like actual works of art, like famous works of art that have been blended and mixed and molded, or fantasy landscapes that have been just blended and mixed until they’re just delightful to look at. They’re just so visually delightful, the things that artists could never have imagined. There’s no way artists could have come up with these things in so many different iterations, and that’s the goal for Joel Simon. He wants people to use this website that is 100% built on biological ideas as a creative discovery tool.

JOEL SIMON: Everyone needs to find inspiration somehow, and also we have our own processes to do that, and so I think a lot of design is inspired by the results of growth and evolution, but I think actually, the process is just as beautiful as the results, because the process gives rise to some of that beauty that might otherwise be impossible to comprehend.

AARON ROWE: So this is something people can try at home.

LUKE GROSKIN: Yeah, absolutely. You can and you definitely should try this at home. After you check out the video, head over to Artbreeder.com, and it’s genuinely fun and actually truly inspiring.

AARON ROWE: OK, that sounds really cool. I’m going to have to try that out.

LUKE GROSKIN: Give it a whirl.

AARON ROWE: Thank you, Luke, SciFi’s Video Producer Luke Groskin.

Copyright © 2021 Science Friday Initiative. All rights reserved. Science Friday transcripts are produced on a tight deadline by 3Play Media. Fidelity to the original aired/published audio or video file might vary, and text might be updated or amended in the future. For the authoritative record of Science Friday’s programming, please visit the original aired/published recording. For terms of use and more information, visit our policies pages at http://www.sciencefriday.com/about/policies/