Generative artificial intelligence (AI) holds promise not only for the future but also for enhancing our understanding of the past. A research team from Google DeepMind, in collaboration with classicists and archaeologists, developed Aeneas, a machine-learning system designed to assist experts in analyzing ancient Latin inscriptions from 7th century BCE to the 8th century CE. Aeneas retrieves relevant textual parallels, offers visual context, and generates speculative text to fill gaps in the inscriptions.
In tests, 90% of historians found Aeneas’ outputs to be beneficial for their research, improving their confidence by 44%. The system excels in identifying overlooked parallels, significantly speeding up preliminary analysis in the challenging field of epigraphy. While Aeneas provides useful starting points, it can’t replace the need for expert validation of the retrieved data. Overall, Aeneas expands research scope, facilitating deeper insights into fragmented historical texts.
