If choosing can be an artistic act, what am I contributing when I select an AI-generated image?
In my latest video, I explore Marcel Duchamp’s readymades and what I understand as “choosing over producing”. His selection and presentation of an existing object challenged the assumption that an artist must physically manufacture an artwork. But how does this relate to selecting generated imagery, particularly when I use it to develop paintings and ceramic forms?
Since 2022, I have experimented with using my own artworks as image prompts. I collage the generated references, make new works from them, then sometimes feed those finished works back into the systems. I’m now in the fourth or fifth round of this process.
My choices run throughout that exchange. Yet the work sometimes doesn’t entirely feel like mine.
Drawing on Saussure and Barthes, I begin by considering how images and objects communicate meaning, and how context shapes interpretation. I then turn to Duchamp’s Fountain and the relationship between selection, recontextualisation and artistic intention.
There is also an entirely imaginary possum exhibition. The system generated the images, the gallery, the catalogue, the merchandise and even my attendance at the opening. The exhibition never happened, but making it gave me something concrete to examine.
The comparison with Duchamp hasn’t resolved my uncertainty. It has helped me locate part of it in the relationship between what I intend, what the system supplies, how I respond and what I subsequently make.
This video forms part of UNICORN VOMIT | Art in the Generative Era, my practice-led, autoethnographic PhD project. Next, I’ll explore appropriation in art and how that history relates to incorporating generated imagery into new works.
Further reading