How today’s technologies from artificial intelligence to virtual reality are changing the way art is made.
At the Museum of Modern Art exhibition “New Order,” art and technology are having a major love-in. The show surveys how the time-honored bond between the two fields has evolved in the new millennium, arraying works that explore and expand the creative possibilities of today’s technologies. There’s a self-playing video game by Ian Cheng, a digital simulation of Jeff Koons’ Rabbit by Mark Leckey, and Tauba Auerbach’s series of 116 3D-printed geometric objects. In these and more, the digital has been made physical; TIFF and AVI files are given tangible form. They are machine-generated realities.
Of course, it’s not as if art and technology are siloed disciplines. For generations, they have been in collaboration and confrontation, pushing and upending the other’s boundaries. “The two are totally embedded and intertwined,” says Michelle Kuo, the exhibition’s organizer and the Marlene Hess Curator of Painting and Sculpture. “Technology is material and formal in a way that we can’t forget. It’s only by really paying attention to these forms and realities of technology and its effects that we can begin to understand how to change it.”
Change is already on its way. Where traditional artists would wield a brush, sculpting knife, or camera, the 21st century affords artists tools from artificial intelligence to 3D printing to virtual reality, enabling them to create in innovative and novel ways. Advancing technologies have dared art to take new risks and art has responded in kind. Once, the interaction between art and technology triggered ethical and moral questions. For example, how can a machine be an artist? Lately though, it’s broaching the richer issue of artistry itself. It asks: can a machine be creative?
When Machines Do Art
It’s an inquiry sparked by the emergence of algorithmic or AI art, the practice of employing code, data, and machine learning systems to generate pieces of art. These Generative Adversarial Networks have been producing such works for a few years now, but it’s only recently that AI art has found itself smack in the middle of the art (and art-buying) world. Last October, an algorithm-generated portrait, “Edmond de Belamy, from La Famille de Belamy,” went under the hammer at Christie’s and sold for a cool $432,500. Created by French collective Obvious, it was the first-ever AI portrait to go on auction, and its sale spawned headlines variously celebrating and worrying about its implications for art-making.
Should artists be worried about their jobs? Not at all, says Obvious’ Gauthier Vernier. “We see AI as a tool to catalyze the artist’s own creativity,” he tells me. “We don't believe that artists’ jobs are threatened, but we believe that a new type of artist could emerge, one that understands and works with AI models.” The products of this union will naturally be curiosities, half-manmade and half-machinemade. After all, in making “Edmond,” the AI model, like the human that built it, acts as producer and creator, significantly blurring the lines between artist and technology. Here and elsewhere, AI doesn’t just facilitate the artist, but also, imitates.
That’s not to say it can create. “Creativity is a notion that is hard to define,” says Gauthier, “so we like to say that an AI can be inventive, in a way that it can invent something new, unique, and relevant.” What these algorithms and machines do, inventively and unpredictably, is generate.
Generative art is not new: art, music, and poetry created by an independent system has been around long before computers existed. In the 1950s, the likes of Vera Molnár and Ellsworth Kelly were devising strict procedures with rules and instructions that determined how their visual art turned out. The ‘60s ushered in computer art pioneers from Frieder Nake to Manfred Mohr to Roman Verostko, who latterly coined the term algorists, as in algorithm artists. And notably, from the ‘70s, musician Brian Eno began experimenting with generative music—what he called “putting in motion something and letting it make the thing for you”—yielding 1991’s Neroli, 2017’s Reflection, and apps like Bloom and Trope.
Eno has not only been visionary and perceptive about the field, but is drawn to the questions it raises: “Who actually composes music like this? Can you describe it as composition exactly when you don’t know what it’s going to be?” At the same time, he remains in thrall to generative music’s promise—likening its vast “multi-centered” nature to the internet, and urging a reconsideration of the artist’s function. “If you move away from the idea of the composer as someone who creates a complete image and then steps back from it,” he’s said, “there’s a different way of composing.”
Making good on that inheritance, today’s generative artists understand how their role is shifting alongside technology. Digital artist Sean Catangui tells me, “With generative art—creating the algorithms that generate art, but also selecting which of the results to present—the ability to create thousands of compositions in less than a second shifts the role of an artist to that of a hybrid artist-editor-curator.” His generative work, such as Line Tree, C-mum, and Randomized Form, emerged from his desire to enhance his ideas through programming. In the process, they gleaned him an appreciation of code and computer as expressive mediums. These projects, he says, “made me feel like the machine was alive.”
And the machine is alive all over Glitch. You’ll find it in other generative artworks such as Matt DesLauriers’ Retro Album Covers, n–schedé’s algorithmic objects and landscapes, and Anders Hoff’s many interactive forms. With these pieces, the artists may have constructed the machines and models, but as they stir and run, it’s their generative efforts that exceed expectations and hold unending surprises. “I might be just as happy to discover some unexpected behaviour and explore that instead,” writes Hoff. “Overall, this means that the process is continuous, exploratory, and never really complete.”
In the lo-fi, early days of computer art, practitioners endeavored to define and describe what exactly it was they were doing. Just observe how the 1976 anthology Artist and Computer offers a trove of phrases such as “creative processing” (from Mohr) or “scientific aesthetics” (Hiroshi Kawano). Whatever they were, all pretty much suggested that the alliance between art and science works both ways. If technology is made a part of art, art might also be embedded into technology to realize what Kawano termed “the science of human art.”
That script, then, might itself serve as an art form. Chad Weinard, Mellon Manager of Digital Initiatives at Williams College Museum of Art, dubs it “code as its own form of creative writing.” He arrived at that idea after witnessing poet Lillian-Yvonne Bertram’s adaptation of Allison Knowles’ computer poem “A House of Dust.” Originally built with FORTRAN language, the piece is a constantly shifting entity, its lines randomly generated and connected. Weinard responded with his own iteration, populated with collection data from the museum.
Representing an interaction between data, algorithms, and chance, his quatrains are also woven with a collected history and a history of collecting. “I’ve been fascinated with the accumulated textures, biases, artifacts and quirks of that data as it’s been passed on over the years,” he says, “and thinking about what legacy we’ll leave in the data for others to find in the future.”
And future humans will find not only art in our code, but art criticism, particularly in Omayeli Arenyeka’s creation, Art Connoisseur. The Twitter bot raids Artsy for images of art, then generates commentary on each with hit-or-miss accuracy and a snarky, knowingly pretentious tone. “Ooh, it’s minimalism without mortality,” is just one brassy example. Though somewhat tongue-in-cheek, the app underscores the abiding gap between art and its interpretation, if not the pitfalls of allowing a machine to evaluate art.
But Is It Art? 🤔
So how should we deal with such machine-generated realities? Generative and AI art are awesome to behold, but neither are intended to be measurements of any machine’s creativity. Rather, as Kawano once wrote, their immense potential and varied products offer only an “artificial creativity.” Perhaps computer-generated works do more by simply broaching and challenging notions of creativity and authorship, presenting a tension between artist and tool that resolves nothing, but offers us the opportunity to interrogate everything. Which, really, is the point of art anyway.
Just like the open-ended nature of generative art, the ever-evolving relationship between art and technology can only continue to astonish as it brings forth yet more revelatory artifacts, arguments, and possibilities. “Technology can change us, the way we think and look and relate to each other,” says Weinard. “I hope art can make technology more self-conscious, mindful of the future, more personal and textured, more humane, more reflective, more vulnerable, more inclusive, more creative.” That exchange bodes well for the future of both fields. All we have to do is set it in motion.