Russian AI Learns to Read Ancient Egyptian Hieroglyphs
A new artificial intelligence model trained on the world’s largest database of Egyptian texts can now recognize and translate ancient hieroglyphs—offering a powerful tool for archaeologists, students, and museum visitors.
Researchers from AIRI Institute, the Institute for System Programming of the Russian Academy of Sciences, and ITMO University have developed an AI system capable of translating texts written in ancient Egyptian hieroglyphs. The model was trained using the Thesaurus Linguae Aegyptiae, the world’s most comprehensive collection of ancient Egyptian texts paired with translations.
Testing took place in two stages: first, using automated evaluation metrics, and then through expert review by Egyptologists from the Higher School of Economics. Scholars note that ancient Egyptian script poses an extraordinary challenge for machines due to visually similar and context-dependent symbols.
From Research Labs to Museum Galleries
The new technology could soon power augmented reality apps that help museum visitors instantly translate hieroglyphs on artifacts. For researchers, it promises to simplify the deciphering and interpretation of newly discovered inscriptions. Egyptology students, meanwhile, could use it to accelerate their study of ancient languages through accurate and accessible translations.
Access to the system will be granted to qualified researchers upon request.
