Researchers from Ludwig Maximilian University and the University of Baghdad have successfully reconstructed a 3,000-year-old Babylonian poem known as the Hymn of Babylon, which had been lost for over 2,000 years.

The hymn, praising Babylon and its god Marduk, was last studied around 100 BC and has been reconstructed from 30 clay fragments using advanced AI techniques.

A specialized AI program was utilized to analyze and match cuneiform text fragments, indicating that they belonged to a single text.

The reconstruction method involved n-gram matching, vocabulary overlapping, and searching for longest common strings of text, with the project drawing on resources from the Electronic Babylonian Library Platform containing 1,402 manuscripts.

The researchers have restored about two-thirds of the hymn's original length, revealing a lost classic of Mesopotamian literature.

The hymn highlights the importance of Babylon's natural resources and beauty, and includes themes of social justice, such as the acceptance of foreigners and support for the poor.

It describes a verdant paradise and a flourishing Babylon, emphasizing its cultural and spiritual significance.

The city is characterized as socially charitable, protecting the weak and supporting the poor, reflecting a bittersweet memory of Babylon's past glories.

The researchers suggest the likely author was part of Babylon's priestly class, as the poem describes priests as the 'free citizens' of Babylon.

The poem's structure suggests it was authored by a single individual, distinct from composite works like the Epic of Gilgamesh.

The complete work is published in the journal Iraq by Jiménez and Anmar A Fadhil.

The use of AI in reconstructing ancient texts is becoming increasingly common, with researchers advocating for improved computational tools for ancient Near Eastern studies.