A pot : experiments with clay + gen AI
digital ceramics, 2024 - 2025.
A pot is a collaborative project with ceramicist Anna Andersen that uses generative AI as a design tool in ceramic practice. We wanted to know whether AI could add something to our process without undermining the hands-on, material qualities that define ceramics, and what AI tools could actually produce for something as specific and physical as a pot. We began this project hoping adding generative AI tools into our belts would expand the way we make design choices, in form and in process. In this work we treat AI as a support tool, not a replacement for human creativity or craftsmanship.
In the studio, we generated and evaluated hundreds of outputs from two early open-source text-to-3D generators, point-E and shap-E, keeping the forms we saw potential in. The tools were imprecise and inaccurate, but that helped more than it hurt. What came back was often far from what either of us would have sketched or modeled, and that distance added to the process; the rough output from the AI tools left room for us to design on top of it, or alongside it, and sometimes despite it. Each sketch (a 3D model) became a starting point that we altered, remixed, and interpreted through our own design sensibilities and understanding of clay. We then translated digital to physical in the workshop, through a combination of clay 3D printing and hand sculpting.
The project produced a series of pots, but its more lasting result was a way of working; a studio process that folds generative AI into ceramic practice, which we've since carried into other projects. Our making stayed where it always was, in our hands, with AI acting as a generative starting point we could creatively push against.
A pot has recieved development support from Nationalbankens Jubilaeumsfond.
In the studio, we generated and evaluated hundreds of outputs from two early open-source text-to-3D generators, point-E and shap-E, keeping the forms we saw potential in. The tools were imprecise and inaccurate, but that helped more than it hurt. What came back was often far from what either of us would have sketched or modeled, and that distance added to the process; the rough output from the AI tools left room for us to design on top of it, or alongside it, and sometimes despite it. Each sketch (a 3D model) became a starting point that we altered, remixed, and interpreted through our own design sensibilities and understanding of clay. We then translated digital to physical in the workshop, through a combination of clay 3D printing and hand sculpting.
The project produced a series of pots, but its more lasting result was a way of working; a studio process that folds generative AI into ceramic practice, which we've since carried into other projects. Our making stayed where it always was, in our hands, with AI acting as a generative starting point we could creatively push against.
A pot has recieved development support from Nationalbankens Jubilaeumsfond.
3D-printed and hand sculpted red clay, various dimensions (largest is 35 x 30 x 40 cm).
A pot with handles, 3D-printed and hand sculpted red clay.
Vase with necks, 3D-printed and hand sculpted red clay.
Double Amphora, 3D-printed and hand sculpted red clay.
Detail, Vase with necks.
AI-generated model of “a pot” - what we consider an initial “digital sketch.”
My sketches for a pot.
First version of Double Amphora.
AI-generated model of “a pot”.
AI-generated model of “a pot”.
AI-generated model of “a pot”.
Two of our first AI-generated 3D models.
Glazing one of the models, 3D-printed in red clay.
Beaker, 3D-printed red clay, slip wash, glaze.