Through the advances of machine learning and computational capacities, machine vision moved from a scientific endeavor into a broad spectrum of domains. Consumer goods such as Apple’s Face ID, the FaceApp, or Instagram face filters shaped popular culture over the last decade. Similar algorithms currently transform mass surveillance and the tracking of unwanted individuals and groups. From terrorists to protester tracking in Hong Kong. Soon satellite imagery will, on a relentless global scale, have enough resolution to trace faces and number plates. In the 21st century, it is not an all-encompassing god watching us but the vision of machines.
This project maps four contemporary machine vision APIs, replacing digital photography with the learning outcomes of computational systems, creating thick Evidentiary Realism mappings, to reveal the racialism, sexism, and prejudice of these systems.
The header image gives a glance into the reshaping of these systems. There are five men in the image and 15 women. Still, Microsoft recognizes two men and one woman. Imagga’s second-best guess for a tag is ‘men.’