Deep Hysteria, AI is an old sexist


For many centuries hysteria (a term deriving from the Greek hysteron, uterus) was considered a pathology, found by medical and mental diagnosis, to which a vast number of female imbalances can be traced back; the scientific (or fake) assumption was based on considering an innate female predisposition towards an unstable, anxious and nervous emotional state. Although this diagnosis was officially disavowed in the 1980s, the concept of “hysteria” has tenaciously remained in vogue, and stereotypes of women as nervous, fearful and uncertain continue to influence the way they are perceived and treated. The artist Amy Alexander, who throughout her career has highlighted the fragility of the boundaries between the media and the surrounding reality, addresses mental laziness, the mother of prejudices, in the version of her implemented in artificial intelligence. The work Deep Hysteria is a collection of portraits of non-existent people, but rather twins generated through machine learning trained on frames of unaware Youtubers and unsuccessful vloggers. The models generated and declined according to gender and expression are subsequently read by another software which has the task of interpreting them in detail based on the recognized facial expression. Since these algorithms are in turn trained on data chosen and structured by humans, they inevitably reflect the same biases. Therefore, faced with a practically neutral expression, the female twin figures, compared to the male ones, are interpreted more as disgusted, unhappy, uncertain, scared or with various similar states of mind, far from the interpretation of the almost identical almost identical male figures. unilaterally judge a state of calm. In these aesthetic and political results of her research, the artist highlights not only the limits (and potential) of AI but their indissoluble link with the limits deeply rooted in our culture. Benedetta Sabatini