Artificial intelligence has become an increasingly vital tool in reading ancient manuscripts that archaeologists and historians have long struggled to decipher. A new era has begun, especially for manuscripts that are worn, incomplete, encrypted, or extremely difficult to distinguish with the human eye.
Teams working in this field have combined artificial intelligence with machine learning and historical research to analyze documents from the Middle Ages. These documents include diplomatic correspondence, forgotten love letters, and politically charged texts.
Decoding a 408-Page Manuscript
Using one of the developed methods, researchers working on Latin manuscripts were able to decode a 408-page document. This work, long held in the Vatican Library, had remained difficult to read for years.
The manuscript contains more than 34 mysterious symbols and includes some Arabic words. Analysis revealed that the text contains medical recipes and information about treatment methods of the period.
Digital Breakthrough on Burnt Parchments
Another project using this technology focused on Greek parchments damaged by the eruption of Mount Vesuvius in the ancient city of Herculaneum near Naples. Since these parchments could not be physically opened, digital methods were employed.
Techniques such as 3D X-ray tomography, digital slicing, and surface analysis were used to separate the layers. Then, artificial intelligence examined ink traces and letter shapes to reconstruct the missing parts.
Learning Letters and Writing Styles
Researchers are also training models capable of recognizing handwriting from different periods. This allows the systems to distinguish not only individual letters but also the writing habits of specific eras.
Experts emphasize that archives contain millions of manuscript pages, making manual examination nearly impossible. Artificial intelligence enables faster scanning of these vast archives and re-evaluation of previously unreadable texts.
New Access to Historical Texts
These technologies can reveal hidden political relationships, private letters, and details of daily life in documents once considered unreadable. Thus, artificial intelligence becomes a critical support tool not only for modern data analysis but also for historical research.
The core principle of the developed systems is to complete missing letters, words, and sentences with the highest probability based on existing data. This provides historians with a significant advantage in understanding especially damaged texts.
