Original works on paper translated through AI into artificial visual syntax and extracted text.

This series begins with my original works on paper: fragile image-fields made from stains, botanical traces, colour bleeds, marks, fragments, and private asemic notations. I treat each work as a source field — not a finished image to be improved, but a surface already containing a latent grammar.

The AI transformation does not decorate the original. It reads it under pressure. Marks, stains, fragments, and colour fields are reorganized through repetition, mutation, spacing, interruption, clustering, failed alignment, and diagrammatic drift. The result is a second state of the work: an artificial syntax extracted from the image.

Each set contains three parts:

Source Image — the original human work on paper.
Artificial Syntax — the AI-assisted visual translation.
Extraction — a text-based reading of what the transformed image appears to know.

The method reverses the usual direction of AI image-making. I am not using language to generate images. I am using AI to generate language-like structure from images.

The original work becomes a field of evidence.
The AI transformation extracts a diagrammatic language hidden inside the marks.
The final text turns that visual language into a poetic proposition.

Together, the five sets form a speculative archive: field notes from an unknown ecology, where matter, memory, rupture, light, and scattered evidence begin to organize themselves before language fully appears.

IMAGES BECOMING LANGUAGE