A key concept in the creativity literature is Boden’s conceptual spaces. These spaces represent the possible ideas in a creative field. Latent spaces, on the other hand, are a key concept in generative machine learning and creative artificial intelligence. Generative models build latent spaces as maps of meaning and possibilities, where they encode the training data in a semantically meaningful compressed representation, and fill it with new unobserved possibilities. Generation is latent space sampling. I propose there is a correspondence between the operations that can be performed in latent spaces and [[Boden's three types of creativity]]