The Normalizing Machine, an interactive installation developed by Israeli artist Mushon Zer-Aviv in collaboration with the software developers Dan Stavy and Eran Weissenstern, deals with visual notions of “normality” and the ways in which bias is inscribed into and reinforced by algorithms and machine learning systems. While being captured on camera, visitors have to decide from a collection of previously photographed visitor portraits which ones seem more “normal” to them. The datasets assembled in this way are evaluated in order to generate an algorithmic image of “normality”. Zer-Aviv thus tracks face recognition techniques of the 21st century back to practices of facial measurement misused for propagandistic purposes under the Nazi regime as well as to the forensic image practices emerging as early as the 19th century. The Normalizing Machine examines how we perceive “normality” today, questioning whether we can do so beyond subjective categories – as well as the role of photographic technologies as supposedly “objective” techniques with regard to mechanisms of normalisation.
More by Mushon Zer-Aviv: mushon.com/tnm