Behind a locked door in the British Museum, London, Irving Finkel opens a drawer and pulls out a clay tablet imprinted with the tiny characters of the world’s oldest written language. Finkel is one of only a handful of people who can read this long-dead script, known as cuneiform, fluently. Now, thanks to developments in artificial intelligence (AI), computers are being trained to read and translate cuneiform, putting fragmented tablets back together to recreate ancient libraries and even predict missing text.
Cuneiform, the world’s oldest known writing system, was used by the Sumerian people in Mesopotamia around 6000 years ago. It started as a way to keep records of transactions and evolved into a tool for linguistic expression. Cuneiform tablets are now key sources of information about the ancient civilizations of Mesopotamia.
Efforts are being made to translate and decipher cuneiform texts using AI. The Electronic Babylonian Literature project, led by Enrique Jiménez, is using algorithms to connect fragments of tablets and make translations. Another project, the Machine Translation and Automated Analysis of Cuneiform Languages, developed by Heather Baker and coordinated by Émilie Pagé-Perron, is experimenting with algorithms to translate Sumerian words into English with an accuracy of 95%. Additionally, researchers have used AI to predict missing text on cuneiform fragments, and computer vision systems have been trained to recognize cuneiform signs.
The ultimate goal is to link sign recognition systems with modern language translation systems, enabling instant translation of cuneiform tablets by taking a picture with a smartphone. However, a challenge lies in the scarcity of transliterated or translated cuneiform texts in digital databases. Efforts are being made to digitize more cuneiform texts and make them available for research.
