ALGORITHM WATCH:
New study revolutionizes Dead Sea Scrolls dating, might rewrite Israel’s history. Trailblazing interdisciplinary system, combining AI and radiocarbon dating, indicates the precious artifacts may have been written decades – or even centuries – earlier than previously believed (Rossella Tercatin, Times of Israel).
Whereas until now scholars have largely dated the texts based on their paleography — the shape of their lettering — the interdisciplinary research team started its study by carbon dating pieces of parchment from 30 of the scrolls held by the Israel Antiquities Authority.
Once these dates were procured, high-resolution images of the Hebrew and Aramaic lettering from the securely dated scrolls were fed into an AI data-prediction model whimsically named “Enoch.”
The Enoch model, using geometry-based character shape analysis to learn the paleographic characteristics of the timestamped scrolls, could then extrapolate their paleographic shapes and date other scrolls that have not yet undergone destructive radiocarbon testing.
The system was used to investigate high-resolution images of an additional 135 scrolls. When checked by human paleographers, Enoch’s suggested dating was found to be 79% accurate.
The underlying, open-access PLOS One article:
Dating ancient manuscripts using radiocarbon and AI-based writing style analysis
Mladen Popović , Maruf A. Dhali, Lambert Schomaker, Johannes van der Plicht, Kaare Lund Rasmussen, Jacopo La Nasa, Ilaria Degano, Maria Perla Colombini, Eibert Tigchelaar
Published: June 4, 2025
https://doi.org/10.1371/journal.pone.0323185
Abstract
Determining by means of palaeography the chronology of ancient handwritten manuscripts such as the Dead Sea Scrolls is essential for reconstructing the evolution of ideas, but there is an almost complete lack of date-bearing manuscripts. To overcome this problem, we present Enoch, an AI-based date-prediction model, trained on the basis of 24 14C-dated scroll samples. By applying Bayesian ridge regression on angular and allographic writing style feature vectors, Enoch could predict 14C-based dates with varied mean absolute errors (MAEs) of 27.9 to 30.7 years. In order to explore the viability of the character-shape based dating approach, the trained Enoch model then computed date predictions for 135 non-dated scrolls, aligning with 79% in palaeographic post-hoc evaluation. The 14C ranges and Enoch’s style-based predictions are often older than traditionally assumed palaeographic estimates, leading to a new chronology of the scrolls and the re-dating of ancient Jewish key texts that contribute to current debates on Jewish and Christian origins.
The article is quite technical, with a dozen appendices filling it out.
Live Science has an article by Ben Turner which includes some critical commentary by epigrapher Christopher Rollston: AI analysis suggests Dead Sea Scrolls are older than scientists thought, but not all experts are convinced. An AI analysis of the Dead Sea Scrolls, which include texts from the Hebrew Bible, could mean they were composed earlier than experts thought.
Professor Rollston is property cautious about the results and properly criticises some overly enthusiastic rhetoric about the research.
The most entertaining result of the research:
Nonetheless, Enoch also corroborates earlier paleography, notably for a scroll titled 4Q114, which contains three chapters from the Book of Daniel. Analysts initially estimated 4Q114's writing to have been inked during the height of the Maccabee uprising in 165 B.C. (a part of the Hanukkah story) due to its description of Antiochus IV's desecration of the Second Temple in Jerusalem. The AI model's estimate also falls within this range, between 230 B.C. and 160 B.C.
I can't wait to see how this will be deployed by the early daters of the Book of Daniel.
If the paleographic profile of Second Temple-era Hebrew needs to be adjusted backwards by some decades, that could have historical implications. But let's see how well the research holds up. Enoch disagrees with expert human paleographers 79% of the time. That is, the paleographers thought Enoch gave unrealistically high or low dates about a fifth of the time. And Enoch has only an 85% overlap with the C-14 probability distributions. Neither sounds all that impressive to me, but the samples for the latter are small and the implications of the former are unclear.
I can't pretend to be able to follow the detailed mathematical data presented in the article. I leave it for others to comment on that.
We'll see. Watch this space.
PRE-POST UPDATE: I was about to post this when I found Brent Nongbri's post over at Variant Readings: New Radiocarbon Analysis of the Dead Sea Scrolls. He has some useful comments.
Visit PaleoJudaica daily for the latest news on ancient Judaism and the biblical world.