You feel the subtle effects of algorithms while using digital platforms: Spotify automatically plays another song based on what you already like; Instagram shows you the stories first from the accounts you interact with most often; and TikTok, dispensing with agency entirely, just gives you a feed of videos “For You,” no choice about who to follow required. Algorithms are designed so that you don’t necessarily recognize their effects and can’t always tell whether or not they’re modifying your behavior. A new body of work by the interdisciplinary artist and technology activist Trevor Paglen—on view at Pace Gallery’s London venue, with a virtual version online—attempts to visualize their workings.
“Bloom” is a series of high-resolution photographs of flowering trees. The sprays of blossoms are tinted different colors in variegated sections, a slightly nauseating spectrum of reds, yellows, blues, and purples. The colors are the biggest sign that something inhuman has happened: they don’t seem to follow a single logic and their arrangements are too granular to have been executed by hand. As Paglen explains in a video published by Pace, the colors have been assigned by machine-learning algorithms developed by his studio that dissect the images’ textures and spatial arrangements, then apply colors to mark differences. Flowers might stay bright white while the trees’ leaves and branches recede into blues. Looking at the images means trying to decode what the computer was evaluating when adding color.
Flowers are a perennial artistic subject, from the Dutch Baroque memento mori that Paglen references in the video to Andy Warhol’s screen prints. But his visualize how a machine perceives an image. The algorithms interpret no symbolism; there’s no ephemerality or tragedy latent to a springtime blossom. The colors emerge from a mathematical process that could be applied to any other image. The elegiac quality of the series comes from the contrast between the content of the images, familiar to human viewers, and the coldness of the machine’s gaze. We don’t really know what it’s looking for, or at.
Paglen’s recent work, both at Pace and in a concurrent exhibition at the Carnegie Museum of Art, evokes the uncanniness that we feel when using Spotify, Facebook, or Tinder. These platforms purport to calculate our judgements and tastes and then replicate them, serving us our own desires so quickly that we don’t have time to consider how well our identities are being reflected by the algorithms’ decisions. Over the past decade, since he earned a PhD in geography in 2008 from the University of California at Berkeley, Paglen has become famous for using his practice to reveal things that are hidden, making media headlines as much as exhibitions. He moves between formats—photography, collage, renderings, and installations of technological devices—to expose contemporary artifacts like the physical cables that undergird the Internet and souvenir badges from classified Pentagon programs. In recent years he has shifted his attention to artificial intelligence, exploring how machine vision is shaping our perception of the world.
“Bloom” shows that beauty can’t be automated—at least, not by our the technology we currently have. More than a series of visual alignments or colors, beauty lies in our memories of the world, the connection of a flower to the experience of spring inevitably passing. Algorithms lack any understanding of this context; they can only approximate it.
In his “CLOUD” series (2019), Paglen uses algorithms to analyze transcendental photos of the sky; he has continued exploring this technique using the mountainous landscapes in the American West, as seen in the Carnegie exhibition. He applies calculations like Hough Circle Transform, first introduced in 1962 to detect circles in images, and then retains the results on the print so that the viewer knows what the machine has seen: thin white circular outlines with dots at the center identify patterns that the human eye would otherwise pass over. The algorithmic lines recall the jokey meme in which the golden ratio is superimposed on any image and always fits something, like Donald Trump’s hair. Paglen’s series appears ominous—machines attempt to perceive beauty by reducing it to straight lines and perfect shapes—but it’s also a little goofy. The patterns don’t change our understanding of the photographs, and the photographs don’t educate us about the algorithms. They function as illustrations.
Paglen tends to hide his critical epiphanies in sumptuous visuals. Viewers may get lost in color or pattern and turn away after a few seconds. Paglen’s activist bent—the artist as investigative journalist or social educator—competes with his urge to make compelling objects. In the best examples, like the “Bloom” series, these goals merge. Art history meets the technological filter through which we now experience much of visual culture, via iPhone cameras, Instagram posts, and TikTok feeds. Once we learn to recognize the influence of algorithms, Paglen hopes, we might figure out how to counter it and reclaim some of the humanity of our vision.
Trevor Paglen filters photos of blooming trees through bespoke computational processes.Read MoreArt in America, Reviews, artificial intelligence, machine vision algorithm, Trevor PaglenARTnews.comRead More