Designed by artist Trevor Paglen and Microsoft researcher Kate Crawford, ImageNet Roulette—which is on exhibition now at the Fondazione Prada, Milan—applies labels to user-submitted photos by sourcing its identification system from the original Image Net, a Stanford-backed project that assisted AI in identifying structures, objects and even people. Employees from across the US were rewarded with minimal amounts of money for each photo they assigned a label to and trained the AI to identify all “types” of individuals. From “losers,” “unsuccessful persons,” “wrongdoers,” “dweebs,” “millionaires,” and more, these identifiers feed the AI their biases and beliefs. The art project works to exploit these biases and disconcert those on the wrong end of the algorithm, ultimately resulting in skepticism of current AI systems. “The way we classify images is a product of our worldview,” Paglen tells the New York Times. “Any kind of classification system is always going to reflect the values of the person doing the classifying.” Read more there.