Using AI to Translate Ancient History

Cuneiform, the world’s earliest writing system, is now being translated with the help of AI. The ancient writing system consisted of patterns of wedge shapes pressed into clay tablets.

Deciphering these tablets has proven fiendishly difficult for a multitude of reasons. The writing system has been used to create documents in several languages; there was no standardized “spelling” of words, over 900 characters, and no consistent spacing.

Additionally, tablets crack and have missing chunks; pieces of a tablet may be spread around the world, hidden away in museum drawers for a hundred years, with no way of knowing how to put them together. At this time, there are only about 75 people in the world who are fluent readers of cuneiform.  

Today, a global digital library of tablet fragments is being meticulously constructed—a testament to the collective efforts of researchers and historians. Researchers are leveraging the power of AI to piece together tablet fragments, as AI not only facilitates relatively quick translations of tablet writings but also predicts missing words.

The digitization process has evolved beyond mere 2D image capture, now encompassing the creation of 3D tablet models. This advancement allows for enhanced deciphering and increased image quality, furthering our understanding of this ancient script and our history.  

-----

About Neoscholar: Neoscholar partners with world-leading researchers to create research-driven courses and mentor students through meaningful academic research journeys.