Tech boosters and doomers alike wonder when A.I. will be truly be sentient, able to think or feel. Pierre Huyghe asks a less predictable question: What is machinelike about human beings? Reflexes, impulses, routines: His show at Marian Goodman Gallery in Lower Manhattan, titled “In Imaginal,” hints at how alien so-called artificial intelligence really is — and, on reflection, how mysterious we are to ourselves.
In Huyghe’s 2024 video “Camata,” installed at Goodman, the camera pans across cracking bones in a picturesque desert. This skeleton is the scene’s most human presence. Soon, a robotic arm enters the frame, gripping a turquoise stone; an autonomous camera whirs and focuses; a motorized reflector adjusts the light. “Camata” was filmed by a hybrid crew of A.I.-guided and human-operated robots, staked out around the remains of an unknown young man — likely a soldier from a 19th-century war — found in Chile’s Atacama Desert. In what is meant to be a funerary ritual, the robotic cameras spend as much time filming one another as they do examining the man’s rotting shoes or curled hand.
“Camata” is a forlorn and affective artwork, and a brutally crisp picture of human-A.I. interaction. An algorithm edits the film in real time. The software’s motivation is arcane. The work is constantly changing, with no beginning or end.
Huyghe (pronounced weeg), a lauded French artist, is known for his striking environments blurring boundaries of art, nature and technology. Since the 1990s he has made a name for himself by “collaborating” with nonhumans. He’s given a crab a gold mask for a shell, dyed the leg of a dog named Human pink, and attached a living beehive to the head of a nude statue. His current show at Goodman marks the U.S. debut of works, including “Camata,” which premiered last year during the Venice Biennale, offsite at the Punta della Dogana, a contemporary art museum within a maritime customs complex. It demonstrates the ways Huyghe has incorporated A.I. models into his explorations of inhumanness.
The gallery at Goodman is dark and cavernous. Just seven pieces — comprising two videos, four sculptures and three masks — are spread across two floors. In an upstairs room, dimly lit in red, the only work is the startling sight of a person crouching in the corner with a glowing plastic shell covering their face. At seemingly random intervals, the mask — part of a work titled “Idiom” — blurts out nonsense speech generated by machine learning, a series of trills, yeows and slurps.
The effect is alienating. Sometimes the performer seems to track your footsteps. A news release claims the masks “collect imperceptible information.” But when I attempted to make contact they didn’t react. This is not a conversation piece; it’s an impossibly obscure monologue. The masks speak when nobody else is in the room. On one visit, a mask sat on the floor without its wearer, chattering away.
I’m not sure Huyghe’s work is about communication at all. He seems more interested in mutual misunderstanding. His show isn’t about us connecting with robots (any more than his previous work endeared us to crabs or bees) but appreciating the profound gulf between two ways of seeing the world.
Some of Huyghe’s earliest machine-learning works, dating to 2018, approach self-expression with a similar dryness. Huyghe has explored combinations of biofeedback and A.I. models. In some versions, the brainwaves of human volunteers shape the picture; in others, the imagery is purportedly modulated by human cancer cells growing in an incubator.
The second video at Goodman, “Annlee – UUmwelt” (umwelt is a German word for how an organism senses its environment) resembles a waterlogged supercut of wildlife documentaries and pet portraits, smears of green and beige, with darker spots suggesting eyes and noses. The footage supposedly reconstructs mental images generated by participants who were asked to imagine Annlee, an anime character that Huyghe and the artist Philippe Parreno bought the rights to in 1999. (A facial recognition camera attached to the screen tweaks the video sequencing in real time.)
The Annlee character is meant to have no past and no story — implying a certain purity to the participants’ thoughts. Of course, short of nirvana, that’s unlikely. Under their sheen of scientific precision, Huyghe’s brain-A.I. experiments are tongue-in-cheek shortcuts through the artistic process. It’s as if, by the miracles of modern tech, an artist can skip the torture of emoting through paint and clay and plug their subconscious directly into a screen.
You could see Huyghe’s show as a flat-footed attempt at A.I. mysticism. But it’s more interesting to think of this work as intentionally desolate. In the cloister of a gallery, we’re tempted to assign this languid landscape film or jittering impressionistic video a degree of soul. The truth is it doesn’t matter — to the software — if the input is human or not.
“Camata” — the word may refer to acorns fallen prematurely from an oak tree, and perhaps to the young fallen soldier — has a clear precedent in “La Région Centrale,” an experimental film from 1971 by Michael Snow, in which a camera clamped to a robotic arm in a rocky wilderness executes preprogrammed pans, loop-de-loops, and inversions. It makes for airsick viewing. Huyghe’s video is a contemporary attempt to evacuate the artist’s intentions from the art. You’re left to make your own meaning out of what remains.
We’re not told who the man in “Camata” was. No one seems worried about piecing together his life. Not Huyghe, and certainly not the robot filmmakers. They grind on like an automated paparazzi, assigned to obsess over the human form, busy at their peaceful and arbitrary tasks long after their subject has died. It’s a bracingly bleak picture. And a clear-eyed vision of A.I.
Pierre Huyghe, “In Imaginal”
Through June 21, Marian Goodman Gallery, 385 Broadway, Manhattan; 212-977-7160, mariangoodman.com.
The post Pierre Huyghe’s Bracing Dark Mirror of A.I. Has Its U.S. Debut appeared first on New York Times.
