Abstract
This thesis is researching computer vision and algorithmic images grounded in Trevor Paglen`s series Adversarially Evolved Hallucinations from his exhibition A Study of Invisible Images at Metro Pictures in 2017. Computer vision is the field of study that seeks to learn computers to analyze and interpret images and other visual input through algorithms and partly autonomous learning processes. To produce the series, Paglen has together with software developers created a computer vision system that can learn from the information it is given, recognize the objects it is trained on, and express itself in terms of images. These works are manifestations of the computer`s invisible operations in a form visible to humans, serving as physical objects that open up for speculation in how computer vision function. Paglen is through the exploration of this system exhibiting the alien and uncanny world of these seeing machines, but also how they are interrelated with humans and structures of power. It is an artistic engagement with a technology that is shaping how we see, act, and organize ourselves. The series is giving a glimpse into how these machines perceive the world, and how they're interpreting images in a way that is radically different from humans. The works are strange encounters with an unfamiliar form of vision that is increasingly operating in our everyday lives while remaining largely invisible. Through a selection of images from the Adversarially Evolved Hallucinations, this thesis strives to understand the aesthetic and visual relations of the new image technologies associated with computer vision, algorithms, and artificial intelligence. In an interdisciplinary manner with a broad range of theory from media-based and visual studies, this thesis is investigating how the changing status of images, autonomous vision systems, power relations, and the conception of the human is presented through these works. The Adversarially Evolved Hallucinations open up engaging explorations of the non-human agency of the machines and their understanding of images and serve as an invitation to think of images and vision beyond the scope of the human senses.