Fully Automated Award-Winning Photography
Dries Depoorter and Max Pinckers’s Trophy Camera v0.9
Text by Marco De Mutiis
First published in NUMMER 10: Post-Photography, Lucerne University of Applied Sciences and Arts (Hochschule Luzern), edited by Wolfgang Brückle and Salvatore Vitale, 2021
Trophy Camera v0.9 is not a normal camera. It pre- sents itself as a harmless yellow plaything, yet it has an ominous red switch and a small display that gives the warning: «camera is ready». The apparatus has no viewfinder or screen. Nor does it need them, for Trophy Camera v0.9 is not meant to be operated by a photographer in the usual way. Its casing contains a small AI powered camera that has been programmed to take nothing but ‹winning› photos. The artists behind Trophy Cam- era v0.9, Dries Depoorter and Max Pinckers, have used machine learning algorithms to train their software using a dataset based on the award- winning ‹World Press Photo of the Year› from 1955 to the present. The camera compares its own live image to the World Press Photo archive, from which it has ‹learnt› to define an award-winning picture. If it finds an image with at least 90% of the award-winning qualities, it uploads that image to the dedicated trophy.camera web- site. Anything less is instantly deleted.
Processes that apply machine vision to the world of images in this way are becoming increasingly popular. From number plate recognition systems for detecting speeding vehicles to more complex ways of teaching computers to see and identify objects and patterns, machine vision is having a profound effect on the three-way relationship be- tween camera, photographer and subject.
Trophy Camera v0.9 provokes reflection and discussion of these issues. It feels at once like a satire and a dystopian vision of the future of pho- tography, but it also raises questions about the tropes of photojournalism and the gamification of photography, which are fueled by the computa- tional analysis of visual representation. Trophy Camera v0.9 asks us to imagine a world without photographers, a world in which photojournalistic tropes and photographic ‹success› can be defined and outsourced to machines. It asks whether such tropes are already embedded in the way we expe- rience the world, namely through exploitative im- ages of mourning women and innocent children, images that reduce the complex themes of conflict and violence to simple but compelling symbols. It reminds us, finally, that awards and grants already define what winning images are according to their own lenses, cultural biases and image politics, while likes and follows already serve as algorithmic metrics of success in the attention economy of the corporate online platform.