Sexual Data

Abstract

In their controversial study, Yilun Wang and Michal Kosinski claimed to have trained an artificial neural network to determine a person’s sexual orientation based on their facial features. This kind of research can be strongly criticized for relying on outdated theories and potentially fueling surveillance capitalism. However, by focusing on textual information, criticisms of this study overlook the importance of the images used in such a scientific project. In this article I propose that the use of artistic research-creation can be valuable in questioning scientific images and the ideological assumptions they reinforce. My own project Sexual Data, for example, is a research-creation work that uses Wang and Kosinski’s algorithmic images in a way that diverges from their original scientific context. The artistic operation allows us to appreciate that image beyond its scientific illustrative function, amplifying the possible meanings of the image.