Once the sacred protection of the magnifying glass and guessing experts, the Dead Sea Scroll was finally influenced by machine learning.
In new articles in journals PLOS one Researchers at the University of Groningen published Wednesday that merged artificial intelligence and carbon dating and found that many of the reels were older than previously estimated scholars. It seems that some people can go back to the times of the Bible’s authors themselves, not centuries later.
The traditional schedule is based primarily on handwriting analysis and carbon testing tradeoffs and now looks optimistic. We now learn that early dating work was misrepresented by the application of castor oil – an unexpected effect of trying to make the manuscript readable in the 1950s that disrupted radiocarbon results.
Mladen Popović is the principal investigator and professor in Groningen, and his team cleaned up the samples and then dated the samples again. They then trained an AI model that playfully named Enoch in the Bible, said to walk with God, and in the process learned some tips to analyze the ink pattern throughout the scroll. After testing, the date Enoke produced a date that matched the corrected carbon reading, usually with higher accuracy.
I found it wasn’t very small. A fragment from Daniel’s book, long considered later copies, now seems to be as modern as the so-called author himself. Meanwhile, the writing styles previously thought belonging to different eras – Compensation and Herodian scripts – proved that the time spent simultaneously was far more than expected. As usual, history refuses to be neat.
It is worth noting that the AI approach does not require destructive sampling of traditional radiocarbon dating needs, which is an advantage when dealing with the remaining 1,000+ undated rolls.
Nevertheless, some scholars warn of constraints. After all, the date of radiocarbon is parchment, not ink, and like any machine, the AI model is limited by the quality of the data fed. But even cautious experts acknowledge that these findings may reassess when and where these reels were produced. As Professor Joan Taylor of King’s College London pointed out guardiandata show that many reels predated Koumran’s careers – a polite way to say they are unlikely to write them there.
In short, the world’s oldest theological archives may end up being the focus, not through intuition, but through algorithms.
