AI Analysis Creates Dating Amgiguity with Dead Sea Scrolls Before now, one of the few ways to date the Dead Sea Scrolls was through palaeography (the study of handwriting) and radiocarbon dating. These methods place the age of the scrolls somewhere between the 3rd century B.C.E. and the 2nd century C.E. But a new AI date-prediction model may provide a more accurate date for the famous Dead Sea Scrolls. After using this new model, called Enoch, researchers have determined that the Dead Sea Scrolls may actually be older than previously thought. The findings, published in PLOS One, also indicate that two biblical scroll fragments date back to the same time as their presumed authors. By using radiocarbon dating, palaeography, and a machine learning model, researchers working on the European Research Council project, The Hands That Wrote the Bible, narrowed down the age of the Dead Sea Scrolls to a more precise estimate. The team combined data from 24 Dead Sea Scrolls and the machine learning model to apply a Bayesian ridge regression method to the scrolls. From there, the team trained the model for date prediction. Enoch stems from BiNet, an in-house deep neural network that analyzes handwritten ink patterns. Now, Enoch can examine ink at a microscopic level through its curvature (called textural) and character shapes (called allographic). This data can provide analysis that standard palaeography can’t. It can also predict radiocarbon dating results by more or less 30 years. AI and the Dead Sea Scrolls
