Close
Close

ESSAYS Mousse 72

Anthropy: What the Post-Human Looks Like

by Dieter Roelstraete

 

April 26, 2020, in the Upper Midwest: I am celebrating my son’s third birthday—along with Ludwig Wittgenstein’s one-hundred-and-thirty-first—in my mother-in-law’s backyard somewhere blissfully people-free and undistinguished north of Chicago. Public life in the state of Illinois has been in full-on lockdown mode since March 17, meaning that I have not set foot inside a gallery or museum in forty days. I also have not been inside an artist’s studio since then, nor have I been in an artist’s company. I have, in other words, not savored the pleasure of the physical (let alone social) experience of art in a month and a half—surely the longest I’ve ever had to go, professionally speaking, without my proverbial daily dosage of the stuff. I have not been in the presence of objects for weeks—only images, screens, a generalized state of absence rather than presence.

 

Looking at art in an effort to remember why I’m doing what I’m doing has become indistinguishable from the generic routine of “research”—that is, endless hours spent down the internet’s warren of rabbit holes for some vaguely investigative purpose or other. (Leafing through books offers some succor and escape, probably because they are comparatively so physical and present.) Daily gallery notices of online exhibitions or other appropriately distanced, virtual content—an artist’s favored sourdough recipe, a bedroom techno playlist—have only helped to confirm a long-held suspicion that art and the internet are somehow incompatible, irreconcilable. Mutually exclusive entities, “a clear-cut case of either/or.”1 Indeed, online exhibitions strike me right now as about as desirable, nourishing, and plainly oxymoronic as virtual food. I’d rather sit here and wait for the world (and all the art in it) to open up for business again. And, sure, content myself with Google in the meantime.

Ironically enough, much of the aforementioned “research” is at present being performed in the service of an exhibition project (and accompanying graduate course, taught online) addressing the question of art and agency in the age of algorithms. It is my modest curatorial contribution to the expanding debate on the relationship between art and Artificial Intelligence. The exhibition will inevitably be a primarily screen-based affair, richer in images than objects—the kind of paradox-ridden undertaking that raises the question of whether it should unfold in a traditional museum space at all. (Of course it should, if only so we could at least look at it together.2) The project’s working title is After You, conjuring images of a fully automated post-human future familiar to readers of books like Alan Weisman’s The World without Us (2007), Jan Zalasiewicz’s The Earth after Us (2008), Elizabeth Kolbert’s The Sixth Extinction (2014), and, most poignantly of all perhaps, David Joselit’s After Art (2013).3

The core questions posed by After You are well enough known: Can AI produce art, and can it do so without artists? Can computers be programmed to appreciate great art, and can they do so—both the programming and appreciating parts—without us? Does the rise of a certain kind of algorithmic machinery spell the certain death of a certain kind of art? It is telling, of course, that these important questions are currently being raised almost exclusively on the outer edges of the tech world, with all of the aesthetic biases that this myopia inevitably entails. It is in the pages of Wired, and not those of the mainstream contemporary art press, that one is most likely to encounter them.

One particular oddity that keeps surprising me, in seeing these questions asked, is the curious fact that the presumed artistic prowess (or promise) of AI is so often expressed in its command of the human figure, and that the preferred mode of manifesting this mastery, more often than not, is through paint, or the mediating imagery of painting. Indeed, a random search of the coupling of “AI” and “art” is certain to yield unending streams of both middling faux-abstract visages and close-up pictures of half-empty paint tubes, canvases, brushes. One article titled “If an AI Creates a Work of Art, Who Owns the Copyright?,” published on the World Economic Forum website (of all places) in August 2017, features a stock photograph of a messy painter’s working surface that, apart from an actual artist’s hand, really only seems to be missing cigarette stubs in an ashtray and a thumbed glass of booze.4 One inevitably wonders whether the titular question shouldn’t be: “If an AI creates a work of art, what are the chances it will be a painting, painted using paint?” (The answer: zero?) Or another case in point: an article by the same author, published on another website, titled “If an AI Creates a Work of Art, Who Owns the Rights to It?,” which is accompanied by a photograph of a robotic arm drawing a human face.5 Shouldn’t here, too, the question rather be something along the lines of: “If an AI creates a work of art, what are the chances that it will draw a human face or figure?” (The answer: low?) Why, indeed, would a robotic arm directed by a computer program bother to draw a human likeness? Does it not seem more reasonable to expect a drawing robot to draw a likeness of itself—another robot? a motherboard? electrical circuitry? an endless tapestry of mathematical formulas? Is that what the quest for the holy grail of aesthetically advanced robotics is about: machines that can produce yet more images of ourselves?

Nowhere has the puzzling irony of this anthropic principle been on more spectacular display in recent years than in the well-publicized case of the French artist-programmer collective Obvious. They made headlines in late 2018 when their AI-produced print6 sold for a record-setting $430,000 at auction (it was initially estimated to sell for $7,000 to $10,000). It was easily the most expensive work of computer-generated art ever marketed. The (gold-framed) piece in question is a portrait titled Edmond de Belamy (2018), representing the smoky, smudged likeness of an eighteenth-century nobleman, a member of the fictional Belamy family, which has since received the Obvious treatment in full. The portraits were created using a machine-learning framework known as GAN, short for generative adversarial network. In effect the algorithm combs through thousands of images fed to it to “learn” how to churn out images of its own. And this, of course, is the crux of my argument: What are images of its own? What is the point of this revolutionary episode in the conquest of art by technology—a glorious chapter in the unfolding narrative of the rise of the machines—if all it amounts to is a return to the old order of the most canonical form, namely painting, serving the most canonical of purposes, namely portraiture? Shouldn’t the world’s most expensive computer-generated work of art be a more truthful visual reflection of its program’s inner workings? Shouldn’t we be allowed to expect that the first major, widely reported achievement of AI in the realm of art looks, well, less human? Or is this whole galling episode just a painful reminder of the power of the art market in shaping art history—that money talks only to paintings (or, more importantly still, things meant to look like paintings), and preferably to paintings of people at that, regardless of who, or what, made them? Here we are, on the threshold of the anxiously awaited post-human, and whom should we encounter on the other side but a machine programmed to picture humans—and using the most antediluvian of techniques to do so.

The world’s most expensive computer-generated work of art turns out, somewhat soberingly (reassuringly?), to be a “painting” of a nonexistent French nobleman. This twisting tale inevitably brings to mind a little-known chapter from the history of computer graphics, centered, unsurprisingly, on a girlie pic. Sometime in 1956, an anonymous IBM employee used a $230 million military computer called SAGE (short for semi-automatic ground environment, the largest such machine ever built) to draw the curving outlines of a pinup on a glowing cathode ray tube screen, creating what one chronicler has called “the world’s earliest known figurative computer art, and quite possibly the first image of a human being on a computer screen.”7 The lure of the loins, in other words. And a frequent contender for the title of most widely circulated early artwork made using a computer is likewise a female nude—yet another human figure all too familiar from thousands of years of exhaustingly anthropocentric art history. The work in question is titled Computer Nude (Studies in Perception I) and was first composed in 1967 by Leon Harmon, a cognitive neuroscientist, and Ken Knowlton, a computer engineer at AT&T’s Bell Labs, on the basis of a photograph of the choreographer and dancer Deborah Hay. 

Meanwhile, on the other side of the Atlantic, the first real strides in real computer art were being made by the little-known Hungarian artist Vera Molnár, whose work did not resort to human visages, female bodies, and/or the authorial illusion of gestural painting. She chose to face the brave new world and promise of post-human abstraction instead. Now aged ninety-six, Molnár, a cornerstone in my ongoing After You research (others include Beryl Korot and Sol LeWitt), continues plotting unperturbed, the subject in recent years of some long-overdue institutional recognition for her groundbreaking contribution to the phenomenon of algorithmic art. A postwar immigrant to Paris when la ville lumiére was still hanging on to its status as a world art capital, Molnár was a cofounder of GRAV, the Groupe de Recherche d’Art Visuel, a cornerstone of early Op and kinetic art.

And sure enough, so blissfully, consequentially devoid of the visual noise of figuration, hers is a quintessential art for our pandemic, possibly post-human times: an art in which computers deliver a fully automated Computer World of angular, shape-shifting geometries, grids, and math (and a very crystalline math at that). Zeroes and ones. Think of her work as the not-so-obvious, both in its refusal of digital art’s deeply ingrained anthropic bias, and in its contrarian insistence on the ancient magic of presence—on encountering this emerging Computer World in old-fashioned embodied person. If only to better grasp the irreducible materiality of even the most avowedly virtual of art forms: the printer’s needling of that most hallowed of analog technologies, namely paper.

As I commit these concluding thoughts to real paper in my Zoom-worn head, there is still no certainty as to when galleries and museums will finally reopen their doors to the public in our armpit of the world. Perhaps it can be so arranged that the first artwork we will lay our eyes upon after emerging from the long, dark night of this human-made cataclysm will be a mid-1970s Molnár print—from her Transformations series, perhaps. I can’t think of a more poignant way to return to the physical fact of art in what will doubtlessly be a world altered beyond recognition—and a lot less human indeed. Whatever the case may be, let us look at it together.

 

[1] As I put it myself, back in 2017 in an essay titled “Work/Out,” written for the catalogue accompanying the exhibition The Everywhere Studio at the ICA Miami, which is where I first speculatively proposed this “law.” (I was prompted to do so in part by one critic’s response to documenta 14, the exhibition I was working on at the time, and of which it was said that “one could be forgiven, walking around this documenta, for believing the internet does not exist.”) The aforementioned essay similarly posited “a long-held suspicion concerning the larger relationship between art and the information & communications technology complex—one of deepening disconnect, and possibly mutual exclusion: a clear-cut case of either/or. Perhaps a certain idea of art—one intimately bound up with the ideological subtext of an exhibition such as documenta, and the idea of art that, for better or worse, continues to reign supreme—is facing certain death, in the Hegelian sense of the ‘end of art,’ because it cannot, or can no longer, be reconciled with 24/7 networked life. Perhaps the complete and total digitization and virtualization of culture, economy, politics and society (indeed, of life as such) has made art, in the historical sense of what we have known this noun to denote throughout the modern and post-modern era, superfluous: a historical phenomenon or, in Hegel’s literal formulation, ‘thing of the past.’” Dieter Roelstraete, “Work/Out,” in The Everywhere Studio (Munich: Prestel, 2018): 193–194.
[2] Which returns us to the aforementioned matter of the quintessentially social charge—its fundament, really—of the art experience. The current pandemic has evidently taught us many things about art beyond the obvious point of its status as an “essential business” necessary for the survival of our species. One of them surely is the extent to which the formerly abstract joy of art has been exposed to be a collective affair, dialogical at the very least. To the many attempts at defining art that our past century has seen, the events of 2020 can now add the following: art is what you (turn away from to) talk about—with whoever else is in the room.
[3] Note the Hegelian echo, in Joselit’s title, of the death or end of art.
[4] Robert Hart, “If an AI Creates a Work of Art, Who Owns the Copyright?,” World Economic Forum, August 16, 2017, https://www.weforum.org/agenda/2017/08/if-an-ai-creates-a-work-of-art-who-owns-the-copyright. A comparably titled online article is likewise accompanied by a stock photograph of a painter’s cluttered toolbox (the image is credited to Pixabay): Shelby Rogers, “Who Owns Art Created by Artificial Intelligence?,” Interesting Engineering, https://interestingengineering.com/who-owns-art-created-by-artificial-intelligence. It is of course no coincidence that these articles worrying about copyright issues, which are really financial proprietary issues—if a work of art created by an AI gets sold, who gets paid?—rely on stereotypical imagery of painting as the symbolic stand-in of all art, for it is painting that will forever command the highest prices in the marketplace. Brushwork equals money.
[5] Robert Hart, “If an AI Creates a Work of Art, Who Owns the Rights to It?,” Quartz, August 15, 2017, https://qz.com/1054039/google-deepdream-art-if-an-ai-creates-a-work-of-art-who-owns-the-rights-to-it/. 
[6] Though it was heavily dependent, apparently, on code developed by a nineteen-year-old US student named Robbie Barat.
[7] Benj Edwards, “The Never-Before-Told Story of the World’s First Computer Art (It’s a Sexy Dame),” The Atlantic, January 24, 2013, https://www.theatlantic.com/technology/archive/2013/01/the-never-before-told-story-of-the-worlds-first-computer-art-its-a-sexy-dame/267439/.

 

Dieter Roelstraete is curator at the Neubauer Collegium for Culture and Society at the University of Chicago,
where he also teaches. Recent exhibitions at the Neubauer Collegium gallery have featured the work of Zachary Cahill, Anna Daučíková, Jason Dodge, Assaf Evron, Goshka Macuga, John Preus, R. H. Quaytman, Martha Rosler, David Schutter, and Cecilia Vicuña.

 

Originally published in Mousse 72

 

Related Articles
Mousse 73
Poetic Justice for Yu Man-hon: Cici Wu
(Read more)
ESSAYS
New Pictures for the Old Ceremony: Julien Nguyen’s Floating World
(Read more)
Mousse 73
The Crushing Weight of the Comfortable: Rodney McMillian
(Read more)
Books
Book Reviews by Quinn Latimer
(Read more)
ESSAYS
Ça m’est égal: The Involution of Desire
(Read more)
Mousse 73
The Copper Marble: Jean Katambayi Mukendi
(Read more)