Though it was only a matter of time before automation came for creativity, the advent of AI art generators has been divisive in the creative community. Some artists have embraced the new tech; others believe it’s #notreal. We at WIRED have spent a lot of time sorting out our own reactions. Undeniably, the technology comes with issues—economic, ethical, legal. At the same time, it expands opportunities for art and artistry. WIRED will always come down on the side of the future. That means encouraging the use of new tools, while refusing to allow those tools to replace human creators.
AI image generators don’t just spit out random acts of art. They must be “prompted”—carefully guided by their human users. (The prompt is the real art form, Kevin Kelly argues in a recent essay.) So we asked eight of our favorite artists to re-create their original WIRED covers using these new tools. Judge the results for yourself. —WIRED Art Department
Dan Winters, Photographer“Using the AI was entertaining and somewhat inspirational, especially with complex prompts. I couldn't be happier with the results—this is actually a better cover than the one that ran.”
Prompt Example: “Pixelated taco centered on robins egg blue solid color”
1 / 7
Ian Allen, Photographer“It’s currently not amazing. But it’s also not terrible.”
Prompt Example: “A graphic close photographic portrait of a man wearing google glasses and various HUDs”
1 / 10
Margaret Swart, Design Director“I've been trying to better articulate to myself the differences between prompting DALL-E and working with an illustrator. This is where I landed: Aside from the fact that DALL-E works in a more literal world of "things," it also gave me repurposed versions of the information that I supplied. I think that's the most important aspect. Illustrators think differently than I do. They take the information that I supply, and then they expand on it. They do their own supplementary research, and they offer alternate approaches and unexpected solutions. Some of the DALL-E results were unexpected for sure, but they were based (it appeared) on random data-crunching from the details I provided. Because I wasn't providing the unexpected ideas in the prompts, I didn't get the innovative solutions. AI (or at least this version thereof) has a huge reserve of data, but not the curiosity and the experience to look at a concept from a different POV—an artist's POV. I don't see severing ties with my illustrator database anytime soon.”
Prompt Example: “Type illustrations ‘Less’ ‘Artificial’ ‘More’ ‘Intelligent”
Most PopularBusinessThe End of Airbnb in New York
Amanda Hoover
BusinessThis Is the True Scale of New York’s Airbnb Apocalypse
Amanda Hoover
CultureStarfield Will Be the Meme Game for Decades to Come
Will Bedingfield
GearThe 15 Best Electric Bikes for Every Kind of Ride
Adrienne So
1 / 4
Stephanie Gonot, Photographer“As a photographer and image-lover, it was really fun to type in words in order to create an image, rather than wrangling my gear and props. The control I missed was made up for by the fun surprises. I started this exercise by trying to get as close to my cover image as possible. I got quite close! But as I tried to get even more specific in my description, the result got further and further away from my cover image. Then I tried typing in lines of the cover story and also the initial prompt I received when I was approached about the job. ‘The Care and Feeding of Interplanetary Civilization’ is a pleasing grouping of words, and I got both some really wacky and really beautiful results. I feel like DALL-E could be helpful to photo editors and art directors when coming up with an idea for a photoshoot. It’s nice to be able to create original references for a mood board rather than using images that already exist. I’m hooked!”
Prompt Example: “Mysterious photo of food floating in outerspace through a spaceship window”
1 / 8
Alvaro Dominguez, Illustrator“Sometimes I add a couple of artists in order to get a more random style. Like, I love Robert Irwin and Jonas Wood, two very different artists with nothing in common. So I add to the prompt 'in Jonas Wood style and Robert Irwin style.' This adds a little bit of both in colors and shapes here and there. But as you can see, I added the words ‘Jonas Wood’ to almost every prompt, just to add some warmth and aliveness to a tech image.”
Prompt Example: “3d render of an oil painting of planet earth as a machine at the top of a cliff with a control panel and a sky with clouds in the background, rainbows, pigeons, fireworks, in the style of jonas wood, in the style of robert irwin, in the style of dan flavin”
Most PopularBusinessThe End of Airbnb in New York
Amanda Hoover
BusinessThis Is the True Scale of New York’s Airbnb Apocalypse
Amanda Hoover
CultureStarfield Will Be the Meme Game for Decades to Come
Will Bedingfield
GearThe 15 Best Electric Bikes for Every Kind of Ride
Adrienne So
1 / 14
Jessica Pettway, Photographer“The visual archive DALL-E is pulling from feels like one of the worst archives I’ve ever seen in terms of diversity in visual language and culture. When trying to make my piece, certain words and phrases weren’t returning logical results. For example, when searching 'woman wearing box braids,' the first results were white women in loose braids. When trying to explore the concept of digital blackface, it wasn’t necessary to add ‘caucasian’ or ‘white’ to the search phrases because that was already the assumed default, even when searching items and themes specific to Black culture.”
Prompt Example: “hoop earring,” “mouth smiling wearing grills,” “white hand with super long nails pointing”
Sam Cannon, Photographer“DALL-E has an impressive dataset to pull from, but I did find some pleasure in the fact that, despite the 650 million image/caption pairs it uses to generate images, it wasn’t able to render out a close replica of my cover. While it had no problem creating images with similar elements (needles, hands, pills) and mimicking the overall style (dark environment, harsh light from above), the abstract idea of a hand pushing through a pin screen was outside of its capabilities … for now.”
Prompt Example: “A 3D rendering of a figure with its head in its hands being pierced with thousands of needles, octane, harsh light”
1 / 5
Patrick Savile, Designer and Illustrator“It was kind of hard to get what I originally wanted, so I was shifting the concept around a bit until I could get an image as close to the idea in my head, which turned out to be egg-based.”
Prompt Example: “High detail airbrush painting of an egg cracking open containing clouds, storm clouds in the background, shimmering psychedelic colours, high contrast, masao saito, Michael whelan, hajime sorayama, high resolution”
Most PopularBusinessThe End of Airbnb in New York
Amanda Hoover
BusinessThis Is the True Scale of New York’s Airbnb Apocalypse
Amanda Hoover
CultureStarfield Will Be the Meme Game for Decades to Come
Will Bedingfield
GearThe 15 Best Electric Bikes for Every Kind of Ride
Adrienne So
1 / 13
This article appears in the February issue. Subscribe now.
Let us know what you think about this article. Submit a letter to the editor at mail@wired.com*.