The work Mutant in the Mirror is an auto-theoretical-experiment centred on queerness. It seeks to problematise notions of normativity and naturalisation reproduced in and through artificially intelligent systems. It can be read as a gesture of resistance, a means to rethink encoded meaning beyond binary hegemonic, heteronormative classification and categorisation. The piece poses the question, how might interactions with certain types of AI co-produce selves beyond the gender binary?
The work uses Generative Adversarial Networks (GAN), a generative model using deep learning, in which two neural networks compete, discriminating between real and fake data with each other in order to become more accurate in their predictions – in this case, the generation of alternative physical representation of the self (my/self). The StyleGAN2 model used in this experiment was trained using a small personal dataset of 70 images across different space-time contexts. The 3000+ image(s) and video(s) generated through the experiment function as a mirror, through which I get lost, disoriented and reoriented, turning differently towards myself, my body, and my corporal and subjective experience of the world—becoming-with-AI. The mutant in the mirror functions like a soul, which comes back to itself to know itself, surfacing a new mutant reality. This work, part of a larger research project, aims to contribute to theory and practice, offering insight into a queering of AI.
If you wish to learn more about the work, we invite you to read the paper titled Mutant in the Mirror: Queer becomings with AI.