Training Humans, engineering moral judgments into AI


AI now plays a decisive role in contemporary image making, both as the perceived ‘new dimension’ of visual culture, and as an engine that allows us to edit and produce new parafictions. Artist Trevor Paglen and AI researcher Kate Crawford have produced a critical work on the harvesting of training images for machine learning. In “Training Humans”, exhibited at the Prada Foundation, they show images from the Imagenet platform as they are labeled and categorised by humans in real time. In turn, these industrial ‘engines of seeing’ reveal their own developing moral judgments and biases, which in turn shape our ways of seeing and experiencing the visual world.


Trevor Paglen | Kate Crawford – Training Humans