The Digital Landscape from August 26 to October 5, 2024 at The Stamp Gallery | University of Maryland, College Park | Written by Oliver Foley

Art can be divided into two elementary actions: observation and synthesis. In the creation of a piece of art, the human takes in their surroundings, their own thoughts, their friends; then their neurons disassemble, reassemble, combine, and distort the data which was once an observation. That which is exterior to the artist is digested into a viscous chum within the mind, a fuel which can then be channeled through the hands, mouth, or body of the artist into their work. This synthesized work is then re-observed and re-synthesized in the cyclic action of art. 

In his pieces Machines Learn From the River: Submerged Printer 1.1 and 1.2 (2024), the artist Billy Friebele invites non-human agents into his craft. Machines collaborate with Friebele in both steps of the art creation process: a video camera extends his ability to observe, and generative artificial intelligence extends his ability to synthesize. 

The use of tools to extend the senses is a phenomenon nearly as old as humanity itself. We possess the innate ability to expand our bodies through the machine, bringing the inanimate into the animate. Friebele’s Machines Learn From the River take their inspiration from two primary sources: Friebele’s own experiences moving through the Anacostia River, and his video piece River / Printer / Habitat. From the perspective of a fish in murky water, Friebele’s camera moves through the shallow waters of the Anacostia, moving quizzically around a large mass of green printed circuit boards and plastic. The camera circumnavigates the pollutant, a large printer with rusting circuit contacts and exposed innards. Through the camera, Friebele is able to see the printer’s vastness to the fish. The eye is limited by the size and movement of the body, but the camera is not; Friebele is able to exit his human perspective and reflect on the proportionality and appearance of the abandoned underwater machine. 

The video was then re-observed by Friebele and his second non-human accomplice: a Generative Adversarial Network (GAN). Just as Friebele’s neurons melt his senses into a collection of electrical impulses and neural connections, the artificial neurons of the GAN turn images of the underwater into abstract tokens. This digestive step of synthesis occurs entirely outside of our awareness. In fact, our awareness itself is a product of this digestion. The next step is more conspicuous: the human or machine searches for patterns in the data. These patterns can be internal within the data or relate to past observations. 

A GAN actually consists of two neural networks, and can be thought of as two people discussing their observations and trying to find patterns in them. Each machine-mind extends its abilities through the other, just as Friebele extends his own. These neural networks then realize their findings in the form of an image which mimics the patterns they have found. This image lies in the center of each piece of Machines Learn From the River. The image is bisected in two by a fluid-looking substance, the GAN’s understanding of water; green amorphous blobs protrude from the reflective sludge giving the impression of a plant. A far cry from the uncanny AI images which now populate the internet, this GAN’s output is much more impressionistic and chaotic. This lack of realism emphasizes the imperfect subjectivity of the observation-synthesis process.

 

Machines Learn From the River: Submerged Printer 1.1 (2024) by Billy Friebele

 

In the final step in the cyclic process of Machines Learn From the River, Friebele learns from the machines. He takes in the aquatic patterns analyzed and regurgitated by the GAN and applies the same observation-mimicry that the GAN performed on his video. Using paint and natural materials from the Anacostia, Friebele paints the image. Particles of soil intermix with the colors, with a reflectivity imparted by the resin coating. The interpretations of the machines give way to a painting lying in the space between reality and technology. Layered stones and other natural materials pull the patterns back to reality, contextualizing the curves and colors with their real-world counterparts. Each stage of Friebele’s artistic process is apparent, allowing the viewer a glimpse into the internal machinations of both Friebele and his non-human companions. 

Billy Friebele’s work is included in The Digital Landscape at The Stamp Gallery of the University of Maryland, College Park, from August 26 to October 5, 2024. For more information on Billy Friebele, visit https://www.billyfriebele.com/. For more information on The Digital Landscape and related events, visit https://stamp.umd.edu/articles/stamp_gallery_presents_digital_landscape.