Normally, August is a slow month in the art world. But last week’s late-summer tedium was disrupted by a certain Jason M. Allen of Pueblo West, CO, winning in the Digital Arts/Digitally Manipulated Photography division of the Colorado State Fair. Allen’s piece, one of three he entered, was not your typical third-rate painting or photograph submitted to a state fair. It was a lab-printed giclée of an artificial intelligence-generated ‘artwork’ rendered on Midjourney, one of several programs that use verbal prompts to create an image integrated from a database of already-existing artworks. In other words, the piece in question was not visually rendered by its ‘creator,’ whose role was limited to typing words into a text box. The rest of the labor was performed by a bot and a local print shop.
Following the predictable Twitter frenzy, multiple media outlets including The Washington Post, Al Jazeera, CNN, Art News and Artnet reported on the outrage. There was widespread concern that A.I. could render analog art obsolete (or at least cause lethal financial damage to the artists). The other ‘problematic’ aspect of the new medium was the issue of plagiarism. The problem seems endemic to the process of word-to-image ‘artworks’ generated from a cache of manmade paintings, drawings, and photographs. Such works arguably commit artistic larceny under the guise of introducing a democratic, cutting-edge technology. Some critics posed the basic (and legitimate) question of whether A.I.-generated art could even be considered art at all. It is surely significant that the New York Times article, penned by Kevin Roose, was filed not in the Arts section but in Technology.
Roose’s piece ended with a spectacularly provocative utterance by the proud holder of the Blue Ribbon in the Colorado Art Fair Digital Arts division: ‘Art is dead, dude, It’s over. A.I. won. Humans lost.’ This puerility could be easily dismissed. Consider the source. The win was in a single division in a State Fair pavilion—Allen was not even competing against painters, sculptors, or photographers. And those ancient media still stand strong. Even the former Tate Director Sir Nicolas Serota failed to convince the world of conceptual minimalism’s hegemony, which he promoted for almost two decades of his tenure as the Turner Prize jury chair. Could the Colorado Department of Agriculture, under whose aegis the win was granted, possibly be the force to anoint the replacement for ‘human’ art? Probably not.
Allen’s awe at the prospect of ‘some otherworldly force’ miraculously spewing a multitude of images is like the delight of a kindergartener at the transformation of his handprint into a turkey. To judge by the NYT article, Allen is a novice to artmaking. He is, in effect, a kindergartener. So while his enthusiasm for the magic of A.I. is endearing, it does not indicate insight into artmaking. At this level, Midjourney is ‘dope’ to the ‘nifty’ Paint-By-Numbers of yore. Many of the artists that took to Twitter, condemning A.I.-generated art for causing ‘the death of artistry,’ easily recognized Allen as an imposter, rather than a professional colleague. Their sentiment was understandable given that, according to The Strategic Alumni Project, 66% of recent art school graduates now carry substantial student debt. Of course, this debt is only the financial consequence of studying art. It does not begin to account for the years of study, with inevitable emotional setbacks and creative aggravations, involved in taking a degree in fine arts. Having taught art history to graduate and undergraduate students for over fifteen years, I can attest that they have to put a lot of work into understanding the art-historical past.
Finally, Midjourney and other programs that generate A.I.-assisted art configure images from the pool of preexisting artworks sorted by humans into categories that reference subjects, genres, and styles. These categories have been established over decades, and in some cases centuries, by critics, creators and consumers of the arts, so that the imagination of A.I. is necessarily formed and limited by the human imagination. A.I. merely compiles and re-arranges material that could be equally understood, stylistically and art-historically, in human terms. Take Allen’s prize-winning work, for example. The image is in landscape orientation; its focal point is a backlit aperture that takes up approximately one third of the background. The dimly lit middleground contains four figures, all have their faces obscured from the viewer. The foreground is a reflective floor surface that accentuates the central grouping of the two caped figures, flanking a centrally-placed single one in a white tunic. Three additional artificial light sources diffuse a golden glow towards the sides of the composition.
Let’s locate the aesthetic progenitors of this image.
The highbrow French title of the piece (‘Théâtre D’opéra Spatial’), its stagey lighting and pompously posed figures invoke the work of the symbolist painter Gustave Moreau, whose decadent canvases of Salome from the mid-1870s provide the closest visual referent.
Except that Moreau’s work, for all its extravagance, is considerably more subtle and compositionally sophisticated than Allen’s. A closer link is to late-nineteenth-century academic painting by the likes of James Tissot. Unlike Moreau, whose idiosyncratic style was highly original, Tissot amalgamated the achievements of his impressionist and symbolist contemporaries into hackneyed academic formulae of touchless finish and overexplained narratives.
Yet for all the tedium of Tissot’s genre scenes, his paintings function in a way that A.I.-assisted art, at least presently, cannot. It tells its story in human terms, taking into consideration facial expressions and body language. Herein lies one of the major factors preventing A.I. images, no matter how realistic, from functioning as art: while they can generate an image from a bank of other images, they lack the capacity to generate empathy in the viewer. Since empathy is essential function for human social interactions, it is also indispensable in appreciating art. Without the ability to generate empathy, images are not art, they are just random stock pics categorized by subject and style.
So how does real art work? I will use as an example the Hollywood-based, Russian-born artist Lena Moross. Moross’ work reflects her initial academic training at the State Academy of Art in her native St. Petersburg, as well as her graduate studies under John Baldessari and Jonathan Borofsky at the Pasadena Art Center College of Design and the California Institute of the Arts. Her dual artistic heritage also explains the bifurcation in Moross’ oeuvre, which rests somewhere between painting and drawing. Her paintings are done in ink and watercolor, as if they were drawings, are often large-scale, and almost always monumental, regardless of their size. While she works on the scale associated with modern and contemporary art, she also references nineteenth–century artists known for small-scale work, primarily book illustration and magazine caricature. Foremost among them is Jean Ignace Isidore Gérard, who used the pseudonym Grandville. He was first noticed for mocking the rituals of Parisian bourgeois life in his 1829 Les metamorphose du jour, a set of seventy wittily-lithographed figures with human bodies and animal heads going about their daily business.
Grandville’s work, especially his early caricature, is iconic by virtue of its graphic style (close to cartoons) as opposed to any aspirations to realism (close to photography). He presented his figures as types, without lifelike detail. Grandville’s drawings are some of the earliest examples of the cartoon tradition that is still strong, and now the subject of psychologists’ research.
A 2017 study that investigated the correlation between the depiction of characters in cartoons and empathy by eye-tracking and EEG analysis concluded that, when viewers are presented with two different types of empathy-inducing cartoons with similar story plots, one in the iconic style and the other in the realistic, there were no significant differences in empathy factors. Moreover, iconic cartoons, the study said, had the advantage of abstraction during empathy induction. Viewers did not require a detailed presentation of the scene. As long as the characters’ body language and facial expressions were legible, viewers exhibited cognitive responses to emotional scenes. One key element in engaging spectators’ empathy is ‘perspective-taking’—the mental ability to imagine the subjective point of view of the Other. It is incumbent on the artist to find the visual means to induce empathy, particularly perspective-taking. Moross does this masterfully.
Most of her work is about the human condition—the essentials of our existence such as aspiration, conflict, suffering, death. But despite the grandeur of her subject matter, Moross does not cross the line into pathos. Her reflection, and her self-reflection, is subtle and never didactic. Sometimes, she presents the viewers with images that seem to reference her personal relationships, but they are never exclusive, insider conversations. The 2012 Boat for Pierre series was likely dedicated to her late friend and mentor artist Pierre Picot, but the empty boat is there for the boarding, if only one could find a way to hop into it without falling into the water. The images trigger the fear of loneliness—it is instinctive to empathize with a solitary, unmoored boat.
Another 2012 group of paintings, Anno Domini, takes us back some two thousand years, drawing attention to our historical minuteness. Here, a lonely lamb of ivory fleece and bright pink highlights gazes at the spectator. We know how that story ends, so again we submit to momento mori contemplation. The 2013 cycle Furniture, while having nothing to do with the Bible or the ultimate sacrifice, carries the same theme of eventual abandonment. This series is based on sidewalk sightings of discarded couches, some in picturesque red, others in washed-out greens. But even the picturesque ones, such as Red Couch #6, we know, will meet their inevitable and unglamorous demise in the city dump.
Moross cranks up the emotional volume in the works that depict people. These could be divided into two groups: images that show people going about their daily life (Men at Work, 2012, and I Killed My Husband, 2018) and more layered ones like For the LOVE of CARMINE, 2016, Buttnekkid, 2017 and Protagonist, 2020-21. According to a statement on the artist’s website, I Killed My Husband is based on found photographs she stumbled upon in a garage sale. Moross found it peculiar that the pictures showed only women, with ‘no men in sight…. so, [she] made up a story…the women got so sick of domestic chores and went nuts and decided to off with their husbands in a true American gothic way!’ Cleverly, the paintings themselves skip the dramatic and the gory, leaving women to perform those dreaded domestic chores. Moross keeps the story she made up to herself, allowing the viewers to concoct their own narratives.
This is one of her power moves: to preserve the ambiguity of interpretation while providing the viewers with the clues to finish the story. She hooks us by activating our empathy, getting us emotionally invested in the creation of a story. The models of the 2017 Buttnekkid series (some of the work from the series was curated, rather intrepidly given the contemporary #MeToo movement, by Mat Gleason), interact and emote amongst each other. But here too, we must work out the clues on own. There is some major drama taking place, so we worry, we pull back, we breathe a sigh of relief. Empathy makes us get involved, and accept the subjective perspective of the other.
A year before Buttnekkid, Moross created a body of work For the LOVE of CARMINE, about a trans woman, Carmine Messina, who also posed for the series. The artist’s admiration for her model and muse spills out of the paintings in the exuberant play of lines and colors. Moross performs a miracle of sorts, transforming Carmine’s heavy, voluptuous form into an elegant dancer and poseuse. This series is one instance when we catch Moross exercising her own empathy, rather than setting up the usual emotional trap for the viewer. She is all in, celebrating Carmine’s fabulousness.
In Protagonist, her most recent completed series from 2020-21, Moross steps back into the shadows, once again shifting the emotional onus onto the viewer. Here she appropriates an iconic work by the Spanish master Diego Velasquez, citing his 1645 painting Dwarf Sitting on the Floor, that depicts Sebastian de Morra, a courtier of Prince Balthasar Carlos. The usual interpretation of this painting emphasizes its existentialist message: despite his wee stature, subservient status, and the telling fact that he is sitting on the unfurnished floor, de Morra is dignified. As one writer put it, he is ‘not taken in by our courtly ways, our unconsidered and shallow cult of appearances.’ Moross recognizes that de Morra would make the perfect protagonist, so she places him in the midst of the action (something to do with dueling poets circa 1830). There he rests, noble and eternal like a miniscule Buddha, gazing at the viewer and imploring us to get involved, to thwart the impending and profligate duel fatality.
Seen in the context of the recent prize-winning giclée that caused such an outcry, Moross’ work serves as an excellent example of what sets apart human intelligence from an artificial one. While A.I. is capable of concocting hyper-realistic images that have wide appeal, at least for now, it is not capable of empathy, and since empathy is a necessary component of art, the fear that A.I.-generated images can supplant the work of humans, or as Jason Allen succinctly put it, ‘A.I. won. Humans lost.’ is unfounded…
If and when A.I. learns how to induce empathy it will be time to panic.
Special thanks to John Barrymore for his assistance on this article.