Abstract
Edgelands is a speculative body of images that showcases a bespoke approach to machine learning image generation and dataset curation as a new pedagogy for artists. This work explores the increasing tension between the natural world and the infiltration of electronic waste. While 70% of new technology is recyclable, only 30% of it actually gets recycled. As devices increasingly get smaller and more advanced, their ability to be recycled drastically decreases due largely to custom fabrication techniques. This is leading to an enormous amount of material blanketing the surface of the earth and, worse, a culture of hazardous extraction practices in illegal e-waste dumpsites. This body of work investigates the ramifications of our capitalism-driven desire for the newest and best alongside the environmental crisis to which these discarding behaviors lead. Edgelands is a research project into technology using technology. The project speculatively explores this situation through machine learning—“breeding” images of midwestern landscapes with images of illegal e-waste dumpsites in Africa, Asia, and India. The resulting trained neural network hypothesizes a world where the mass of discarded electronics creeps into the periphery of everyday life and occupies the spaces abandoned by previous industries. The output from this newly trained model speculates on what this future might look like should we continue on the current trajectory. The images are simultaneously familiar and foreign, present and future, and encourage viewers to rethink their relationships to technology, devices, and the lifespan of said products.
Presenters
Jonathan HanahanAssociate Professor, College of Art, Washington University in St. Louis, Missouri, United States
Details
Presentation Type
Paper Presentation in a Themed Session
Theme
2024 Special Focus—Images and Imaginaries from Artificial Intelligence
KEYWORDS
Generative Imagery, Dataset Curation, Hacking, Process-Driven Pedagogy, Thick Interfaces