Drawing on theology, Andrea Pinotti’s contribution explores the concept of virtuality and its manifold implications in the age of artificial intelligence and virtual reality. Beginning with the etymological roots of “virtual” in virtus, the author traces a lineage through Scholastic philosophy, Aristotelian metaphysics, and modern thinkers like Deleuze and Lévy. The notion of virtual containment (continentia virtualis) is central, positing that the virtual is not unreal but a realm of latent potentialities awaiting actualisation. This metaphysical framework is then applied to AI, especially in relation to latent space in image generative models, where virtualities (e.g. styles) are harboured and actualised in particular pictures through prompts. Through experiments with text-to-image model DALL·E and referring to Neural Style Transfer, the essay highlights how generative AI seems to presuppose a clean separation between form and content: this leads to a problematic extrapolation of the style as a universal from a given corpus of images, which ends up reviving the ancient Scholastic debate on universals. In the last section, the essay considers the adjective “virtual” as it appears in the strictly technical sense of the term “Virtual Reality” and highlights analogies and differences in AI and VR as technologies of actualising virtualities.