Artificial Intelligence in Medicine: AI Brings to Light Signature Writing of the Dead Sea Scrolls

What do the Great Sea Scrolls have in common with diagnosing disease? This blog takes a detour from Artificial Intelligence (AI) in medicine to a different realm before circling back to clinical use: how many individual scribes contributed to the Dead Sea Scrolls?

The Dead Sea Scrolls

In 1946, a treasure trove of ancient Jewish and Hebrew manuscripts was first discovered by Bedouin shepherds in the Qumran Caves, near the northwest shore of the Dead Sea. The manuscripts were scrolls that were preserved in earthenware jars. They proved be complete manuscripts, unlike the larger regional body of manuscript fragments.

The scrolls were written in Hebrew, with some Aramaic, Latin and Greek. Archaeological dating techniques place the age of the Qumran scrolls from the 3rd century BCE into the 1st century CE. Many scribes during these centuries set stylus to parchment, recording sacred texts that are part of the Hebrew Bible (Old Testament) and other religious writings.

The scrolls are one of the greatest 20th century discoveries related to the ancient Middle East. They are of tremendous historical significance, especially to scholars of the Old Testament. According to The Conversation, “Included among the scrolls are the oldest copies of books in the Hebrew Bible and many other ancient Jewish writings: prayers, commentaries, religious laws, magical and mystical texts. They have shed much new light on the origins of the Bible, Judaism and even Christianity.”

Handwriting reveals unique traits

Who were these scribes? There is general agreement that many, if not all, were members of a religious sect called the Essenes. They lived in organized communes and practiced asceticism. Like the European monks in medieval monasteries, the Essenes wrote and copied manuscripts, thus preserving texts that were important to their history and faith. Sadly, we will never know their personal identities, since they did not sign their work. Or did they?

No two people—not even identical twins—write exactly alike. Each person’s writing is unique, and therefore a telltale clue that sets one person apart from another. Even when people are schooled in uniformity (remember learning to print the letter “a” repeatedly on ruled paper?), many physical and neurobiological factors go into shaping each stroke of pen or pencil. A trained analyst can spot tiny distinctions. The elements of handwriting can be analyzed on at least two levels: textural features of each letter, such as the curve and length of a line, and the gestalt of the entire letter. Thus, it’s possible to identify individual scribes by their unique style, even if we don’t even know their names.

AI distinguishes one scribe from another

Scholars approach the study of the Dead Sea Scrolls on many levels: history, content, meaning, etc. Some have focused on handwriting, which has let to academic dialogue and debate over how many scribes may have contributed to a given scroll. This is a daunting challenge. Take, for example, the Great Isaiah Scroll, written in Hebrew. The Hebrew “a”, or aleph, appears over 5,000 times in the scroll. While the human eye may be able to detect very small differences among a few copies of aleph, analyzing and characterizing such a large data set would be very time-consuming and prone to subjective opinion.

Now, thanks to applying the power of AI, this type of analysis is accurate and efficient. A May 3, 2021 news story relates the breakthrough use of deep learning to crack the identity code embedded in fine details of human handwriting. “Digital imaging makes all sorts of computer calculations possible, at the microlevel of characters, such as measuring curvature (called textural), as well as whole characters (called allographic).”[i]

The project was conceived by a team from the University of Groningen (Netherlands), and involved several steps of developing and training a state-of-the-art artificial neural network based in deep learning, a subset of AI. Once the algorithm recognized the ink markings separate from the background surface, it learned to differentiate the telltale traits of each scribe’s muscle movements. Even scribes who may have been taught to write uniformly have “person-specific” qualities unique to their fine motor coordination. Thus, the team was able to definitively conclude that the Great Isaiah Scroll was begun by one scribe, and completed from roughly midway onward through by another. This suggests that AI can ultimately identify the number of scribes who labored over their scrolls.

In a similar manner, when AI is appropriately trained using a large number of medical images such as pathology slides of cancer cells or MRI scans of prostate tumors, it is able to present radiologists with patient-specific diagnostic information. As I have previously written, this will increase the efficiency and accuracy of disease identification

We have come a long way since the Dead Sea Scrolls were created, but we can apply the same analytic tools to generate individual live-saving treatment plans with the speed needed when time matters most.

NOTE: This content is solely for purposes of information and does not substitute for diagnostic or medical advice. Talk to your doctor if you are experiencing pelvic pain, or have any other health concerns or questions of a personal medical nature.

[i] “Artificial Intelligence Helps Crack the Code of the Dead Sea Scrolls”. University of Groningen, May 3, 2021. https://scitechdaily.com/artificial-intelligence-helps-crack-the-code-of-the-dead-sea-scrolls/amp/