ABOUT
The Trans-Feminist Queer Transformer is a GPT-2 model fine-tuned on datasets of trans-feminist and queer lyrical poetry to produce textual output for musical composition and live performance. More than half of this dataset is taken from my own decade-long career as a singer/songwriter specializing in queer-feminist folk music. This model is currently housed in the lab in which it was created: the BMO Lab for Creative Research in the Arts, Performance, Emerging Technologies, and AI within the University of Toronto's Centre for Drama, Theatre, and Performance Studies.

RATIONALE
The aim of this project is to explore the conceptual-material possibilities and impossibilities, as well as creative and ethical tensions, of conceiving of a practice of collaborative authorship between myself, a human singer/songwriter and artist, and a GPT-2 machine learning system. Some of the questions that guided my work were as follows: How can artistic research methodologically guide my working through complex entanglements of embodiment, authorship, and autonomy in collaborations with artificially intelligent systems? What are the simultaneous possibilities and limits of engaging artificially intelligent models in collaborative creations of knowledge, and to what extent can trans-feminist-queer experiences and desires find resonance with machine generated performance text? What insights would I gain through the process of feeding my own queer and female desire, language, poetry, voice, and experience into a language processing system and asking it to reflect those things anew? To what extent can artificial intelligence allow me to encounter, express, and reach an enlightened understanding of my own desires in the creative process? And, finally, What would it mean to let myself be transformed by AI as a creative technology or collaborator, and not as an extractivist tool of surveillance and control?

DEVELOPMENT HISTORY
This project began with technical support and mentorship from David Rokeby, director of the BMO Lab for Creative Research in the Arts, Performance, Emerging Technologies, and AI—housed within the University of Toronto's Centre for Drama, Theatre, and Performance Studies. This model went through two major rounds of fine-tuning in late 2023 and early 2024, and produced output which I then used to create unique musical and lyrical compositions as "evidence" of human-nonhuman "collaboration." This was not without extreme tension, contradiction, and conceptual-material failure. I am currently in the process of refining a scholarly article about this practice-as-research, which will rigorously reflect on the above.