Visions of the past using Artificial Intelligence
See and Believe
Using Generative Artificial Intelligence models the exhibition team created images of the historic landscapes of Vulci and Roselle. Inserting PaleoBotanic pollen species into text-to-image AI prompts the specific historic ecosystems are presented in realistic, photographic detail.
Example AI Prompt:
The hilly landscape around Vulci was characterized by few trees of common oak (Quercus robur), hazelnut (Corylus avellana), ash (Fraxinus excelsior), hop hornbeam (Ostrya carpinifolia), field maple (Acer campestre), hornbeam (Carpinus betulus), smoothleaf elm (Ulmus minor) and few scattered shrubs of blackthorn (Prunus spinosa) and dog rose (Rosa canina).
Example AI Prompt:
Wet environments in the plain were characterized by few trees of white willow (Salix alba), alder (Alnus glutinosa), smoothleaf elm (Ulmus minor), ash (Fraxinus excelsior), and black elder (Sambucus nigra). Wet meadows played an important role, with Cyperaceae-the sedge family (including some wood clubrush-Scirpus sylvaticus), common reed (Phragmites australis), small reed mace (Typha angustifolia), soft rush (Juncus effusus), and meadowsweet (Filipendula ulmaria). Some hydrophytes, such as duckweed (Lemna minor) and flowering rush (Butomus umbellatus), were also present in the shallow water. )
AI Landscape Gallery
