AI may be the tech of the future, but it’s already coming for the past.
It’s the Hebrew month of Elul, when tradition-minded Jews like me contemplate our own shortcomings in preparation for the new year and Day of Atonement. So I’m here to tell you that, for my sins in the year 5785 soon drawing to a close, I’m writing this month about AI and how it is affecting the field of biblical studies.
In this case, the AI, named Enoch by its developers at the Qumran Institute of the University of Groningen in the Netherlands, looked at “microscopic ink trace patterns such as curvature and character shape” in fragments from the Dead Sea Scrolls — the oldest existing copies of the Hebrew Bible, found in caves near the village of Qumran by the Dead Sea. The Jerusalem Post reported on an article published by PLOS One in June 2025 that said the advance has “profound implications” for the dating of all kinds of ancient manuscripts.
Enoch was able to identify not only the individual scribes who handmade copies of the books of Daniel and Ecclesiastes, but also predict with great accuracy when those copies were made based on the analysis of how Hebrew writing changed between 300–50 BCE. That timeline was in turn based on data from new, highly sensitive radiocarbon dating of the scrolls themselves — a process that the researchers do not employ lightly because they have to destroy samples in order to analyze them.
It turns out that the Book of Daniel in the possession of the Israel Museum is even older than once thought, down to an accuracy of ±30 years instead of the old 150-year range. But more importantly, Enoch is creating a rubric of writing developments by using telltale imperfections that even expert paleographers — ancient handwriting specialists — can’t see. In that way, it works a bit like a lie detector, which tracks differences in skin conductivity and blood pressure that are invisible to the interrogator and not under the subject’s conscious control.
The paleographers in Groningen were good before Enoch came along, but now they have solid evidence that can be expanded to ancient Greek and medieval manuscripts as well, thanks to AI. Undated texts — or even texts that consciously try to sound older than they are, a well-known phenomenon throughout Jewish history — will yield up their secrets, at least in terms of timing.
Another recent article in TheTorah.com uses not AI, but statistical analysis to date the books of the Bible. A Portland University professor of modern Hebrew, an emeritus professor of Statistics at the U. of Illinois and a Hebrew Bible scholar at Emory teamed up to write about the differences between what is known as Classical Biblical Hebrew (CBH) and Late Biblical Hebrew (LBH). They use the example of the words mamlakha and malkhut, both usually translated as “kingdom,” to illuminate the statistical probability that a single scribe switched between the two forms of the word randomly (a decimal point with sixteen zeroes after it) or that they were used by contemporaneous scribes who spoke different dialects (also essentially a statistical impossibility).
This is because some of the earlier books of the Bible — like Samuel and Exodus, written in CBH — contain an overwhelming majority of the uses of mamlakha in the Bible while LBH texts like the Book of Esther (which, incidentally, is the only biblical book not represented among the Dead Sea Scrolls) disproportionately use malkhut. For you stats nerds out there, mamlakha appears in CBH texts 86 out of 92 times, while malkhut is found in LBH stories 51/53 times. The scholarly dividing line between CBH and LBH texts is the Babylonian Exile, 597–538 BCE, so we are talking about a linguistic change that seems to have its roots in the experience of a few generations.
What does all of this mean to me, a historical novelist rewriting the Exodus on the basis of modern scientific research? Only that, like everyone else in 2025, I have to keep an eye out for AI potentially undermining everything I think I know. The Dead Sea Scrolls were written centuries after the events I explore in my book take place, but their analysis by Enoch has just begun to get results. As a writer of fiction, I thrive on the myriad controversies and disagreements in the field of biblical archaeology; I get to pick the most dramatic possible interpretation of those arguments and use it for my own purposes. If an AI like Enoch establishes, for instance, that Mount Sinai wasn’t a volcano or that there was never a red algal bloom in the Nile, my whole current project is cooked.
Luckily for me, I’ve found my artistic mission writing novels inspired by the entirety of the Hebrew Bible, not just the stories of Exodus. And since the Bible happens to be the most inspirational book ever written, in terms of how many lives it has profoundly changed throughout the millennia, I’m pretty sure I’ll be able to pivot.
