Description (in English)
How To Be More Or Less Human examines how computer vision interprets gender, race and class through commercial stock imagery. Visual stereotypes and typologies are highlighted and performed by human and non-human actors to demonstrate how humans are encoded into computer vision algorithms. In attempt to be recognized as a male by image recognition software a performer takes off items of his clothing until the software incorrectly identifies his naked body as female. The performance reveals how gender bias and visual stereotypes influence artificial intelligent software and performs the relationship of training between the human subject and machine learning applications.
Text Source: https://maxdovey.hashbase.io/howtobemoreorless/