DOODLE: make a mark

A synthographer spends a great deal of time asking machines to make images. DOODLE began with a different question: what happens if the machine helps make an instrument instead? Working with ChatGPT, I developed a simple browser-based drawing tool: a blank surface, a restrained palette of colours, four line widths, fill, an eraser, undo, clear and save. Nothing more elaborate seemed necessary. This is not an attempt to imitate professional graphics software, nor is it a demonstration of draughtsmanship. It is an invitation to make a mark on screen. My usual practice begins with language, constructing prompts and allowing an image model to respond. Here that relationship is reversed. The machine helped build the apparatus; the human receives the prompt. Two quiet words — doodle here — are enough to begin.

That reversal changes the question of authorship. The visitor makes the mark, but makes it inside a visual system whose dimensions, colours, behaviours and restrictions have already been decided. There is freedom, but never complete freedom. This is not so far from my usual synthographic practice, where authorship is exercised through direction, constraint, judgement and selection rather than by placing every pixel by hand. DOODLE sits adjacent to synthography rather than pretending to be another form of it. A photographer can make a book, a painter can make an installation and a sculptor can make a drawing. A synthographer can make an image-making instrument. What matters is less the category of the finished object than the continuing interest in images, machines, control, chance and the negotiations between them.

Any doodle made here is not an AI-generated image, and there is no reason to pretend otherwise. It is a human mark made inside an AI-assisted system that I helped to design. After years spent constructing increasingly sophisticated synthetic images, there is something grounding about discovering that, given the digital equivalent of a crayon, I can still struggle to draw a convincing potato. I like that imbalance. Behind every primitive line or field of colour sits a surprisingly complicated technological structure; behind the technology sits the simple pleasure of moving a hand and watching a mark appear. DOODLE deliberately offers less: fewer controls, fewer possibilities, no prompt engineering and no promise of perfection. The machinery recedes, the mark remains, and making an image becomes as simple — and as difficult — as deciding where to begin.