At the Columbia symposium on art, science and AI that I wrote about last month, an artist on the panel asked another panelist, a computer science professor, what he thought about AI being trained on artists’ work, enabling a computer to mimic their style. The professor had just described using AI to create an image for which they didn’t have the budget to hire an artist. He responded, “Artists are trained on artists’ work.” A murmur rippled through the audience. Although his comment missed the point that training a model on an artist’s entire oeuvre and creating mounds of artwork in the same style with the click of a button is kind of horrid, potentially hurting the artist’s ability to sell their own work, he wasn’t wrong.
This topic had already been on my mind because of an exercise I’m eager to try with the scientists I work with, but before I get to that I’d like to explore the issue by grappling with the times I’ve borrowed famous artists’ styles for my own work.
An obvious question is: why are you mimicking the style? In some cases, a client has asked me to, such as this candidate cover art in the style of French avante-gard artist Sonia Delaunay (1885-1979), the first living female painter to have a retrospective at the Louvre.
The client had a specific reason for choosing this artist, and explained it in the submitted cover description. The journal’s editors said that even though it didn’t look like any one of her paintings, the style was too similar. As I wrote about in the September 2023 issue of this newsletter, we were sent back to the drawing board, and this artwork now only hangs in the office of one of the disappointed clients.
Another client asked me to create cover art in the style of Takashi Murakami (1962-). This was dicier because he is still living, but his paintings sell for several millions of dollars apiece. I made the cover below for a highly technical scientific journal, so it seems unlikely it would affect his livelihood. It is also much too different from any of his actual works to risk legal trouble. But, what would he think? I like to think he would say that he was influenced by anime and manga (which he was), and that every artist is influenced by previous ones.
“Good artists copy; great artists steal” is a quote usually attributed to Pablo Picasso but also sometimes to Steve Jobs. I’ve never liked the quote, but I understand the sentiment. Copying is what our 10-year old does when he draws characters from the video game Brawl Stars as impressively faithful representations of what he sees on the screen, then gives them away to his friends. It’s what Hunter S. Thompson did when he retyped large portions of the work of F. Scott Fitzgerald and Ernest Hemingway to get a sense of their rhythm and style. It’s artists training on other artists’ work. The second portion of that quote means taking elements of another artist’s style but making it your own. I considered whether it would bother me if someone used my style. My only somewhat discernible style is probably my proteins, which I would have to admit were inspired at least a dozen years ago by the character Meatwad from Aqua Teen Hunger Force. In this cover art for the Journal of the American Chemical Society, I took elements from the style of Wassily Kandinsky (1866-1944) but made it my own.
Much of Kandinsky’s work is now in the public domain in the US because they were made over 95 years ago. So technically I could have submitted an exact copy of one of his paintings for cover art with some hand-wavy justification for why it describes a mutagenizing chimera. But it’s better that I made this homage, right? Or is it? What if Kandinsky would have been horrified by this inferior piece that undiscerning eyes might have inadvertently attributed to him? What if it damaged his reputation? I could also consider whether it would have offended his sensibilities. Kandinsky didn’t create abstractions of real objects like Picasso and others did, but rather created harmonies of color and form on the canvas, inspired by his love of music. He believed his art should only come from within, from what he called an “inner necessity”. I used his style but not his intention, since my forms are abstracted proteins and DNA. Then again, maybe that is what makes it my own.
The piece below was inspired by advances in gene editing that enable deletion or replacement of long stretches of DNA. No one asked me to draw it in the style of Joan Miró, it’s just how I pictured it when I envisioned the molecular process. Like the Delaunay tribute, it was passed over for cover art so not used for anything, and like Kandinsky’s work, many of his paintings are now in the public domain. If it seems like I’m seeking absolution, it’s worth noting that the AI field is still figuring out the ethics themselves. The generative AI image-creating Midjourney has a pre-set called “Joan Miró style”. But if you ask Adobe Illustrator software to generate “bowl of macaroni in the style of Joan Miró”, it will nobly respond, “One or more words in your prompt violate user guidelines. Please edit and try your prompt again.” I doubt the problem is macaroni.
What about Andy Warhol (1928-1987)? He had his own habit of appropriation, so I don’t feel too bad about this one. I think he would have approved, even if it was used to “sell” science as journal cover art. Unlike Kandinsky, who vehemently opposed materialism, Warhol’s work celebrated materialism and consumerism. Likewise, Murakami (above) entered into a collaboration with Louis Vuitton.
Given the widespread use of the Picasso quote about copying/stealing, drawing on someone’s style for inspiration is a generally accepted practice, considered inevitable, and has even been co-opted by another group by replacing the word artists with designers. From my digging, the general consensus is that it can and should be done, but with proper attribution.
Believing strongly in the benefits of an artistic practice to creativity in science, I began planning an art workshop for the scientists where I work. Some of them are artists themselves, but most don’t identify as such. People often claim, “I can’t draw,” unfairly excluding themselves from a world that is so much bigger than representational art. My hope is that inviting them to make art about their science in an abstract artist’s style might free them to draw and paint without worrying about style or skills. (This is not to say that abstract artists don’t have skill, but that they used it to do the work of developing the style already.) I believe such an activity can help people ponder their scientific work in a new light, question assumptions, and surface new ideas. But I paused to consider the ethical implications posed here. Having mulled it over, I see zero harm in this activity because the purpose is simply the process. Though I do dream of filling a wall or two in our space with these artworks as a perpetual source of inspiration.
Absolutely love your homages to Miro and Kandinsky, Mary. Another great topic well done. Thank you!
"I doubt the problem is macaroni" 😂 I love Miro and I love that the Twin PE paper has been honored by your illustration ala the sensibilities of Miro here.
Hard to know if any of these artists would be flattered or offended (or somewhere in between those two extremes) but I think you're right to factor in both the context and intention of the artist mimicking the original style. It doesn't take anything away from their style or intention if you draw upon it to communicate science or to use it for a different purpose.
If anything, perhaps it's something artists should expect. Once in the "public domain" it will ripple through networks of individuals and their collective imaginations...what comes out the other side as a "copy" is an almost inevitable echo of the audience processing the work. The work, as writers often say, goes on to have a life of its own post publication. Then, does it not belong to us all? These are admittedly late night and only semi coherent ramblings of my brain :) Always a pleasure reading your words and seeing your beautiful art!