NEW YORK, NY.- In 1434, the high-tech medium of oil paint allowed the Flemish master Jan van Eyck to infuse his sumptuous double portrait of the Arnolfinis with astonishing depth. He couldn’t resist showing off a little more: A convex mirror on the back wall contains a tiny self-portrait of the painter at work.

Six centuries later, when the multimedia artist and writer Seth Price includes an illusionistic mirrored sphere in the upper left of “Thought Comes from the Body II,” a big, crackled black and Day-Glo painting on panel, it still signals virtuosity. In the reflection, you can make out two figures — one of them may be Price — crouching over a painting on a studio floor.

Van Eyck had to eke out this illusion by hand. Price used an app. His latest paintings, on view at Petzel in Chelsea through June 3, highlight a question that’s easy to gloss over: Why are today’s technologically attuned artists using advanced software, including AI, to invoke the past?

For two decades, Price — a New York conceptual artist collected by MoMA and the Whitney — has adopted the manufacture and distribution of contemporary art as his subject matter, starting with his 2002 essay-artwork “Dispersion.” Akin to suspending pigment in oils or coating glass with photosensitive silver salts, Price’s paintings since 2020 have combined abstract dashes and pours with ominous portraits, jots of text, and backgrounds resembling notebook or sketchbook pages — with the strikingly contemporary addition of trompe l’oeil chrome tubes and mirrors seemingly poking through their surfaces. To render those perfectly warped reflections, Price photographs a painting, adds shiny objects in 3-D modeling software, then matches those shapes to the physical panel with an industrial printer.

Five of his 11 paintings on view at Petzel incorporate AI-generated imagery, mostly buried in abstract spills and smears. But some have tells. The shaky anatomy and lumpy physics in Price’s “Weken Style,” a demure black and white diptych of a warped group of figures and a studio table cluttered with dreamlike tools, suggest the work of machines, as does their square composition, a format favored by the generative imagemaking AI called DALL-E. Indeed, Price conjured the pictures using AI, printed them “wet” on plastic, then smeared the ink with his fingers, adding an inimitable human touch. (Many of Price’s latest paintings feature another emblem of the artist’s hand: the brush stroke, which artists from Roy Lichtenstein to Laura Owens employ to signal that, even though the technology and the thinking have changed, their work is still Art.)

The evidence of AI in “Danlivin” stands out from the painting’s rings and splashes: the nonsensical phrase “THE TNETES 19989,” the 9’s in different fonts. This is the distinctive garbled diction of image-generators, which imitate the look of words but not necessarily their meaning.

The impulse to use AI to indulge nostalgia is palpable. This March, Boris Eldagsen’s winning image in the Sony World Photography Awards in the Open Competition — surprise! — was AI-generated. It sure looks like a vintage photo, though: a black and white, worn-looking picture of two women, one hunching enigmatically behind the other. There’s even a glow in the upper left corner, like a light leak on film. Eldagsen claims he entered the contest to spark discussion (and subsequently rejected the prestigious prize), while the judges assert that they happily selected the work of an algorithm. (Something’s clearly off: As with so many Photoshop fails, the hands don’t match the torsos. Human anatomy, like words, can be tricky for image-generating AIs.)

The ethics of truth in AI-generated media remain fraught, especially where politics and history are at stake. The analog photographers Herbert Ascherman and Shane Balkowitsch point specifically to the way anthropology-style pictures generated with prompts like “tintype of lost New Mexico tribe circa 1800s” could pollute, and effectively erase, the historical record.




But fretting about the use of AI in contemporary art is like ranting against the mechanical loom. Beyond the fog of novelty, it’s worth asking what kinds of images artists ask of their software, and why.

As Eldagsen was shaking up the traditional photo world, Gagosian mounted a show in March at its Upper East Side outpost in New York by Bennett Miller, better known for directing movies like “Capote” and “Moneyball.” These pictures are the progeny of DALL-E — and yet they affect the wooziness of old photos. The square, sepia images portray Victorian children, vanished Indigenous American chiefs, and Hollywood-type “braves” tumbling from the cliffs. A picture of a tiny white flower on a fingertip is speckled as if printed from a dusty negative — flaws that any photographer would have corrected but which, in the age of AI, are the marks of ersatz authenticity.

The way these AI-generated images mimic “real photographs” is an ironic throwback to the 19th-century academic art establishment, which protested that photos, captured with mechanical and chemical means, were more scientific instruments than artistic media. Some photographers back then reacted with “pictorialism” — a style that imitated painting by downplaying technical precision and dramatizing misty depths, soft focus and moody lighting.

Arguments against AI-generated art sound familiar: “There’s no skill involved, you’ve only pushed a button.” And some AI image-makers, too, have reacted by embracing romantic, retrograde styles, including the murky “past” of pictorialism.

There’s nothing in Eldagsen’s or Miller’s fantasies that couldn’t have been made by maquettes, models, even trusty Photoshop — 10, 20, 50 years ago. Granted, AI imaging programs aren’t good at making new things. By design, they can only play “exquisite corpse” with huge collections of what they’ve been told are faces, or animals, or tintype portraits.

Price does more than imagine alternate, sepia-toned pasts. He incorporates the nostalgic impulse of AI as just one of the cacophony of styles that defines our present moment. In their weird mélange of old and new, printer and hand, Price’s paintings embody the simultaneity that defines “the contemporary.” His paintings describe the anxiety around what makes us human that AI provokes — but they don’t succumb to it.

—

Ardomancer: Through June 3, Petzel Gallery, 520 W. 25th St., Chelsea, 212-680 9467; petzel.com.

This article originally appeared in The New York Times.