ABSTRACT
Emergent Artificial Intelligence (AI) and queerness have an uneasy relationship. AI deployment towards detecting queerness, identifying and isolating, or removing and containing queerness has received a lot of attention as another example of how AI continues in the legacy of technologies that have been weaponised against queer bodies and communities. Critiques and resistances to these deployments often take up the idea of improving the data set, correcting algorithms, or expanding the scope of networks to include more diverse voices. In this chapter, I argue that the production of queerness as contamination or dirty is an ontological prerequisite for the development of contemporary AI systems. I show how the narrative of AI as clean and robust—as a saviour and as an enforcer—requires for the queer body to be contaminated, in order to deviate from its own conditions of leakage and bleeding of data. I propose that Queering AI needs a radical rethinking of some of the fundamental ontologies of AI, thus making way for Queer AI as something that doesn’t just treat queerness as a site of implementation, but as a way of reimagining Queer futures of AI.
