Piecing Together an Ancient Epic Was Slow Work. Until A.I. Got Involved

Scholars have long struggled to reconstruct the Epic of Gilgamesh, one of the oldest literary texts, due to the vast number of cuneiform tablets and the limited number of experts in the field. Since George Smith's discovery of a fragment in 1872, efforts to piece together the epic have continued, but about 30% remains missing. Recently, an artificial intelligence project called Fragmentarium, led by Enrique Jiménez at Ludwig Maximilian University, has significantly accelerated this process. Utilizing machine learning, the team has matched over 1,500 tablet fragments in just six years, revealing new segments and hundreds of missing words and lines. These discoveries have provided deeper insights into the narrative, including details about Gilgamesh's friendship with Enkidu and their quest for immortality. Notable new findings include a prayer from Gilgamesh's mother and the introduction of the word "lavish" in Utnapishtim's dialogue, indicating a sense of guilt. The ongoing work with AI is expected to uncover more connections between ancient writings, and researchers remain optimistic about discovering additional fragments in museums and unexcavated sites.

- The AI project Fragmentarium has accelerated the reconstruction of the Epic of Gilgamesh.

- Over 1,500 tablet fragments have been matched in six years, revealing new insights into the epic.

- Approximately 30% of the Epic of Gilgamesh remains missing.

- New findings include significant details about characters and their relationships.

- Researchers believe more undiscovered fragments exist in museums and archaeological sites.