When generative AI is asked to create an entirely white image, what would it in fact generate? The ostensibly blank white surface is not a pre-established colour, but rather the machine’s projection—generated through latent vector operations, probabilistic computation, and denoising processes.

​

The resulting images are materially instantiated via a nineteenth-century iron-silver photographic process. During exposure and development, the image progressively emerges through the precipitation of metallic silver. Computation is thus transposed into a photographic material procedure, allowing machine vision to inscribe itself into the white.

​

In White Surface, whiteness functions both as a manifestation of otherwise imperceptible computational operations and as a residual trace of the computational environment. At the same time, it remains a surface waiting for future projection.

​

​