Field of Potential

Field of Potential: AI diffusion video installation

Three-part HD video installation, 2025

The installation works with diffusion models to hold the face in a continual state of becoming. In these models, images emerge through a recursive process of noising and denoising, a rhythm of dissolution and formation. Between these two states lies a field of potential: a liminal zone where images hover before they are fully formed.

The installation treats this process not as a technical procedure but as a philosophical act; a refusal of recognition. By suspending images within this threshold, the installation resists the demand for identification, and the normalising pull of the face as a surface for capture, sorting, and control.

Developed through workshops and collective experiments, the installation constructs systems where images never completely arrive. This refusal of fixity connects the technical to the political: the same infrastructures that optimise recognition also reproduce histories of racialisation, visibility, and exclusion. The Field of Potential turns the training loop back on itself, asking whether digital faciality can undo its own logics from within.

Rather than reclaiming the face as essence, the work loosens it, allowing subjectivity to fade into relation, into the flux of becoming.