AI-based image-generative tools significantly enhance creativity and digital reconstruction by accelerating workflows for reconstructing digital heritage. These tools apply to various contexts, such as archaeological sites, transformed urban areas, damaged buildings, unbuilt architecture, and partially lost artworks like mural paintings and frescoes. In reconstructing such works, the process is highly complex and requires scholars with expertise in recognizing and attributing fragments and better-preserved pieces. While it remains necessary to suggest potential lines, figures, or scenes, the outcome often involves multiple possibilities. AI image generators can support these challenging tasks by integrating them into workflows, from producing stochastic results to extending patterns and coloured areas. However, their use, especially via “off-the-shelf” software, introduces two key challenges. The first concerns achieving accurate reconstructions that maximize the AI’s potential while avoiding casual or tentative outputs. This involves addressing limitations in existing AI systems and ensuring the final results are both reliable and time efficient. The second challenge relates to cultural evaluation. The line between AI-assisted reconstruction and creating entirely fake results is delicate. Misuse could exploit public fascination with AI or lead to errors and misinterpretations without expert oversight. This integration is in its early stages, necessitating rigorous testing and exploration. The proposed contribution analyzes the digital reconstruction of medieval frescoes from central Italy, blending traditional methods with significant AI inputs. These case studies highlight the balance between human-driven processes and the acceptance of digital outputs, offering reflections on the evolving relationship between human expertise and AI-driven creativity.
From Art for Industry to Artificial Intelligence, a Complex Balance in a Case from the Centrale Montemartini
This paper explores the relationship between historical artistic production and contemporary AI-based image generation through the case study of an Art Nouveau lamppost at the Centrale Montemartini in Rome. A photogrammetric survey is used to document the artwork, which is then analyzed and translated into textual prompts for various AI image generation systems. The results highlight both the potential and the limitations of current generative AI, particularly in reproducing complex spatial, symbolic, and anatomical features, emphasizing the gap between human artistic intention and algorithmic output.
A Parallel Between Words and Graphics: The Process of Urban Representation Through Verbal Descriptions, from Historical Painters to the Automatically Generated Images by Artificial Intelligence
This paper explores the relationship between verbal descriptions and visual representation of urban environments, comparing historical artistic practices with contemporary AI-based image generation. The study traces how medieval and Renaissance painters constructed cityscapes from symbolic, oral, or textual descriptions, often translating partial knowledge into coherent visual narratives. Through the analysis of selected artworks—such as Sassetta’s City on the Sea, Spinello Aretino’s frescoes, and Pisanello’s St. George and the Princess—the authors highlight recurring compositional structures and symbolic elements in premodern urban imagery (e.g., walled cities, gates, towers, and spatial hierarchies).
The research then develops an experimental framework comparing human and artificial processes: architecture students and experts were asked to draw a city based solely on a textual description, while the same prompt was processed through AI image generators such as Midjourney and DALL·E. The results reveal that both human and AI outputs depend heavily on prior visual knowledge, stylistic conventions, and selective interpretation of textual inputs. However, AI systems tend to emphasize mainstream visual patterns, ignore parts of the prompt, and generate images driven by probabilistic associations rather than structured architectural reasoning. The study concludes that AI image generation is not a replacement for human creativity but a tool that exposes the mechanisms of visual translation, offering new insights into both historical representation practices and contemporary computational creativity.
Fragments of Stories and Arts: Hidden and not so Hidden Stories
Any city with a long and articulated past has buildings, squares and monuments linked to its history, the built heritage is its more evident direct link to the historical and artistic events that characterize the present urban asset. In between this main feature, there is the possibility that a myriad of a minor, medium or minimal sized elements may be present, creating a network of evidence, sometimes diffi-cult to catch, but strongly connected to past events and valuable stories. It brings to light details that are often ignored or misinterpreted because of their historical peculiarities. The present research is focused on a structure based on an AR solution to make these traces in Florence downtown more ac-cessible and discoverable. This paper base is the starting point for a special and fascinating exploration of the Florentine downtown, passing by a series of “secondary” but highly intriguing traces. In addition to the most important places and monuments, known and desired by tourists, there are details and trivia that further enhance the uniqueness of the experience in the historical and cultural city.
When the Real Really Means: VR and AR Experiences in Real Environments
During this past year the Laboratory for eXtended Realities (DIDA–LXR) from the DIDALABS sys-tem at the Department of Architecture at the University of Florence, has experienced a various number of activities. Most of them linked together digital modelling of no longer existing architectures and still in place Built Heritage. Others were aimed to develop an “Augmented Virtual Reality” using specific environments/locations (for example a boat) to enhance the sensations of the user during the experience. Some others were based on direct VR shooting, using advanced panoramic cameras and creating a point of view compliant with the specific impressions that the place should transmit. In the contribution proposed here the AR, VR and XR experiences from this personal research will be presented sharing the specific subjects, the evaluation of usable technologies, the strategy for shoot-ing, survey, processing and post–processing, the dissemination of ideas and the lesson learnt.