Learning to see images the way machines do
“Machine Visions,” multimedia artist Trevor Paglen’s first solo exhibition in Korea, is being held at the Nam June Paik Art Center in Gyeonggi until Feb. 2 as a part of the Nam June Paik Art Center Prize. Paglen was chosen as the winner last year.
The exhibition is fairly compact, with 19 pieces that span across various mediums, including video, installation and photography. Each work touches on different aspects, but the essence lies in one of two things: to show what images machines use to help function the way they are programmed and to show how those technologies could be used against us - to censor our lives and use it against us.
For example, “Tornado (Corpus: Spheres of Hell) Adversarially Evolved Hallucination” (2017) is a picture “painted” by artificial intelligence, after it was educated on what a tornado is, to show directly how machine vision and imagination differ from our own. On the other hand, “NSA-Tapped Fiber Optic Cable Landing Site, Keawaula, Hawaii, United States” (2016) is a picture taken from Hawaii of a seemingly calm beach. In fact, submarine cables lie underground, like the political power that exists beyond our sight. All of Paglen’s works are meant to show that “technology is not neutral” through the idea of vision, according to the artist.
“What I see my job is as an artist is to try to learn how to see what the world looks like at the moment in time that you find yourself within,” Paglen told local press. “When we think about the history of images, it’s very historical. And for all of history, images needed somebody to look at them for them to exist. They didn’t exist without the human to interpret them.
But that’s changing now. With things like computer vision, we can now build machines that look at images for us, and most of the looking at images that is going on in the world is done by machines - not by humans. Factories are looking at images for logistics and quality control, while cities have cameras that can automatically identify who people are.”
BY YOON SO-YEON [firstname.lastname@example.org]