This project is an examination into the classifications that artificial intelligence models and algorithms use to categorize, represent, censor, and suppress human’s digital bodies with a focus on sex-positive and queer bodies.

It pushes back on the rigid categorization imposed by tech companies and explores data labeling as a form of empowerment. It questions why we collect the data points that we do, and explores what it may look like to take a more pluralistic, contextual, and temporal approach to data. The Deviant Dataset asks, what does it mean to exist inside the “error” and what are the larger societal implications of this censorship.