Articles by Lucio Fiorini

The paper proposes a methodological reflection on the use of computational procedures and Generative Artificial Intelligence (AI) for the representation and communication of archaeological heritage, taking the work of Giovanni Battista Piranesi as a case study. Piranesi’s engravings, situated between reality and invention and characterized by a complex interplay of survey, interpretation, and imagination, provide an effective testing ground for experimenting with AI-based digital workflows. The research outlines a streamlined operational procedure that integrates image-to-image transformations with the generation of three-dimensional models, starting from historical iconographic sources. AI is employed not to produce automatic or verified reconstructions, but as a mediating computational tool aimed at regulating the representational process and managing the uncertainty inherent in visual data. The procedure is applied to different categories of urban and architectural subjects, including monumental views, archaeological artifacts, and imaginary spaces. This comparative application enables an analysis of how generative models respond to varying degrees of spatial and informational ambiguity, revealing phenomena such as geometric simplification, selective loss of detail, and hypothetical completion of non-explicit elements. The results highlight the potential of integrating archaeology and computation as a critical framework for interpreting, enhancing, and communicating archaeological representations.