An international team of researchers has taken a fresh look at ancient texts using artificial intelligence, shedding new light on who might have penned parts of the Hebrew Bible. Led by Shira Faigenbaum-Golovin from Duke University, the study focused on the first nine books—known as the Enneateuch—and revealed three distinct scribal traditions through careful analysis of word choices and sentence structures.
The project brought together experts from archaeology, physics, and computer science to examine subtle linguistic cues. The team discovered that while texts like Deuteronomy and the historical books share a lot in common, they stand apart from the priestly writings. Even common words such as “no,” “which,” or “king” carried unique stylistic fingerprints that the AI model was able to detect.
One standout feature of the approach was its transparency. The AI not only sorted chapters by their likely authorship but also explained which words or phrases influenced its decisions. For anyone who’s ever wrestled with ambiguous texts, this clear rationale can be particularly reassuring.
Given the Bible’s long history of edits and revisions, traditional machine learning models typically struggle with such short, complex fragments. To overcome this, the team developed a custom method that zeroed in on sentence patterns and word frequencies—a smart workaround that could be useful for analysing other historical documents as well.
Intriguingly, the study also highlighted differences within the Ark Narrative of the Books of Samuel. While 2 Samuel showed a clear link to a specific scribal tradition, 1 Samuel didn’t match any of the patterns. The researchers even suggested that this method could one day help verify, say, whether a document attributed to Abraham Lincoln is genuine.
This interdisciplinary effort not only deepens our understanding of biblical authorship but also paves the way for future analyses of ancient texts like the Dead Sea Scrolls. It’s a fine example of how blending scientific rigour with humanities can make age-old puzzles a little less mysterious.
