Dead Sea Scrolls Just Decoded by an AI… And It's Far Worse Than We Thought

Summary

The video argues that new imaging and computational techniques are helping researchers recover and study Dead Sea Scroll material, while the manuscripts themselves show that ancient biblical texts existed in multiple forms. Despite its dramatic title, the video stresses that AI did not independently read the scrolls or deliver a verdict about the Bible. Instead, it assisted researchers by revealing text and patterns that were difficult or impossible to detect by eye.

The Cave of Horror discoveries

• In 2021, an Israel Antiquities Authority team descended about 80 metres into the Cave of Horror, in the Nahal Hever cliffs of the Judean Desert, as part of a wider effort to protect caves from looters.
• The cave’s name reflects an earlier discovery: the remains of around 40 people who had taken refuge there during the Bar Kokhba revolt against Rome, between 132 and 136 CE.
• Among the finds were roughly 80 small scroll fragments. Their Greek writing linked them to pieces of a scroll discovered in the region in the 1950s. Researchers identified the old and new fragments as parts of the same physical scroll.
• The text is a Greek translation of the Minor Prophets, including passages from Zechariah and Nahum. According to the video, this version differs in places from the standard Greek Old Testament, apparently reflecting an effort to bring the Greek closer to the Hebrew.
• The scribe wrote the divine name using archaic Hebrew letters rather than translating it into Greek. The video presents this as evidence that scribes made deliberate choices about what to adapt and what to preserve.

How imaging and AI help recover the texts

• Many fragments appear blank to the naked eye. Researchers photographed them under different wavelengths, including infrared and ultraviolet, because ink and parchment reflect light differently.
• Multispectral images revealed writing on at least some fragments that had previously seemed empty.
• Pattern-recognition software can help distinguish faint ink from stains, cracks, and shadows, and identify letter forms. As described in the video, its role is to help recover the raw text, not interpret its meaning.
• Images of the large collection—described as roughly 25,000 fragments—were made available online for researchers to examine.
• Software has also helped match fragments to one another and to previously known scrolls. The video says newer models are being developed to estimate a scroll’s age from its handwriting.

Handwriting analysis and scribal collaboration

• A University of Groningen team used computational analysis to study the Great Isaiah Scroll, a nearly complete copy of Isaiah more than seven metres long.
• Earlier scholars had generally regarded it as the work of one scribe because the handwriting looked consistent.
• The team’s neural network measured subtle features of the Hebrew letter aleph. When it grouped the scroll’s columns by handwriting, they fell into two statistical clusters, with a division near the middle.
• The video presents this as evidence that two scribes copied the scroll and closely imitated one another’s style.

What parchment DNA may reveal

• Researchers can analyze DNA preserved in animal-skin parchment and use it to compare fragments.
• The video says most scrolls were made from sheep, while some were made from cowhide—a material unlikely to have been produced locally in the arid Dead Sea cliffs.
• It also describes two Jeremiah fragments made from different animals and preserving different wording. These findings are presented as evidence that texts from different places or communities may have been gathered together in the caves.

Missing books and differing versions of scripture

• The video notes that many books of the Hebrew Bible appear among the scrolls, but no fragment of Esther has been found. It gives two possible explanations: Esther does not mention God by name, and its Purim festival may not have fitted the community’s calendar. The reason for the book’s absence is unknown.
• More broadly, the scrolls do not all match the later standardized Hebrew Bible. The video says some align with the Masoretic tradition, while others preserve readings associated with the Greek Septuagint or the Samaritan version.
• The video gives examples of differences in wording, verse order, and length, including a shorter version of the David and Goliath story.
• Jeremiah is highlighted as a particularly striking case: two substantially different editions were found among the scrolls, including one described as about one-eighth longer than the other.
• The video’s central historical conclusion is that, before standardization, multiple textual forms could circulate and be treated as scripture. It says the Masoretic text became the standard later, after the Second Temple’s destruction in 70 CE, as competing versions fell out of use.

Overall takeaway

According to the video, the scrolls reveal a human history of copying, translating, editing, preserving, and selecting sacred texts. That history does not by itself settle religious questions, and the narrator explicitly rejects the claim that AI has made a theological judgment. Instead, computational tools are bringing previously unreadable material and overlooked patterns into view, allowing researchers to examine how ancient texts were transmitted and how one version eventually became dominant.

Speakers and sources featured

• Ambrose Wilde — narrator and presenter.
• Oren Ableman — identified as a scroll researcher with the Israel Antiquities Authority; cited regarding the Greek translation.
• Emanuel Tov — identified as an Israeli textual scholar; cited regarding the manuscript evidence and textual traditions.
• Israel Antiquities Authority team and researchers — described as conducting the cave survey and studying the finds.
• University of Groningen research team — credited with the computational handwriting analysis of the Great Isaiah Scroll.
• Other unnamed researchers and scholars — cited in connection with multispectral imaging, parchment DNA, fragment matching, the absence of Esther, and theories about the scroll collection.