How can we create new ways for understanding how AI image-production models inform our experience of the world?