Authorship in the model
Sketch, prompt, selection: where authorship sits in generative images.
Generative models deliver images, but someone decides which one stays. These works start from my own drawings, paintings and prints, and follow what a model makes of them and what is chosen from the result.
-
01
Paintings and graphic work
Graphisch
Painted and graphic works in one gallery: broad colour fields, brush marks, repeated lines and fragments of writing.
Look closer -
02
One sketch, two words, hundreds of iterations
Intergalactic Frequencies
A hand-drawn sketch, a two-word prompt, hundreds of iterations. The text is about bias as signature, selection as authorship, and the latent space expressing itself.
Look closer -
03
A model trained on my own images
TheUnpaintedOnes
A generative series trained on painted and graphic works, accompanied by a synth composition. I select the series from the model's images.
Look closer -
04
Rooms from fragments
Inference Rooms
A print series of rooms built from architectural fragments. The spaces appear organised until you try to follow their structure.
Look closer -
05
The Mechanical Turk
Who Asks the Question
In 1770, a human hidden inside a chess machine defeated Napoleon. In 2026, 360,000 people signed up to work for AI agents. The text is about authorship, invisible labour, and why creativity is selection rather than generation.
Look closer