An international team of researchers is using artificial intelligence to illuminate one of the oldest mysteries in literary history: who wrote the Bible. The sacred text is not the work of a single author but a complex tapestry woven from different oral and written traditions over hundreds of years. In a study published in the journal PLOS One, scholars combined statistical analysis with biblical expertise to create a new tool to distinguish between the unique writing styles of the Bible’s various contributors, offering a new, quantitative method for addressing long-standing debates about its origins.

The team’s method focuses on subtle, often unconscious, linguistic habits that can serve as an author’s fingerprint. They developed a novel AI-based statistical model to analyze word frequencies and patterns in the Enneateuch, the first nine books of the Hebrew Bible. To train their model, the researchers first fed it chapters with well-established authorship, categorized by biblical experts into three major scribal traditions, or “corpora.” These included the oldest layers of Deuteronomy (D), the historical narratives in Joshua through Kings known as the Deuteronomistic History (DtrH), and the law-focused Priestly writings (P). “We found that each group of authors has a different style — surprisingly, even regarding simple and common words such as ‘no,’ ‘which,’ or ‘king,'” said Thomas Römer, a biblical scholar at the Collège de France and a member of the research team.

With the model trained, the researchers tested its ability to attribute authorship to biblical chapters with heavily disputed origins. The analysis first confirmed what many scholars already believed: the writing styles of the Deuteronomy and Deuteronomistic History corpora are far more similar to each other than they are to the Priestly writings. More revealingly, the model identified distinct authors for texts previously considered a single narrative. For example, it found that the two parts of the Ark Narrative, a story spanning 1 and 2 Samuel, were likely written by different hands, with the chapter in 2 Samuel showing a strong stylistic connection to the Deuteronomistic History. Crucially, the AI can explain its reasoning. “One of the main advantages of the method is its ability to explain the results of the analysis — that is, to specify the words or phrases that led to allocating a given chapter to a particular writing style,” said researcher Alon Kipnis of Reichman University.

This new approach provides a powerful and objective tool that complements traditional methods of biblical scholarship, which rely on linguistic, historical, and archaeological criteria. The researchers believe their methodology has applications far beyond ancient religious texts. The same AI-driven analysis could be used to verify the authenticity of other historical documents, from the Dead Sea Scrolls to letters attributed to figures like Abraham Lincoln, by identifying the unique stylistic signature of their authors. “It’s such a unique collaboration between science and the humanities,” said Shira Faigenbaum-Golovin, an assistant research professor of Mathematics at Duke University who helped lead the study. “It’s a surprising symbiosis, and I’m lucky to work with people who use innovative research to push boundaries.”

References

• Duke University. (2025, June 5). AI reveals hidden language patterns and likely authorship in the Bible. Phys.Org; Duke University. https://phys.org/news/2025-06-ai-reveals-hidden-language-patterns.html
• Faigenbaum-Golovin, S., Kipnis, A., Bühler, A., Piasetzky, E., Römer, T., & Finkelstein, I. (2025). Critical biblical studies via word frequency analysis: Unveiling text authorship. PLOS One, 20(6), e0322905. https://doi.org/10.1371/journal.pone.0322905