📜 AI Reads Roman Scroll for First Time in 2,000 Years — No Unrolling Required
- NewBits Media
- Jun 28
- 3 min read
It survived a volcanic eruption. Got buried. Carbonized.
Now, for the first time in history, AI reads Roman scroll without ever touching it — and unlocks a message from the ancient world.
When Mount Vesuvius erupted in A.D. 79, it destroyed Pompeii and nearby Herculaneum. But in that devastation, something remarkable was preserved: the only intact ancient library ever discovered, deep beneath the ruins of a luxurious Roman villa.
Archaeologists uncovered 1,800 tightly rolled papyrus scrolls in the 1700s — but they were so fragile, attempts to open them often destroyed them. Many were thrown out or literally turned to ash.
⚙️ Then Came the Tech Revolution
Computer scientist Brent Seales of the University of Kentucky asked a simple question:
“If we can see inside the human body with technology, why can’t we see inside these scrolls?”
So began a fusion of particle physics and artificial intelligence:
Using a particle accelerator in England, researchers generated a light source 10 billion times brighter than the sun to scan the scrolls non-invasively.
AI was trained to detect faint traces of ancient ink, even where the naked eye sees nothing.
🧠 Enter: The Vesuvius Challenge
To push progress further, Seales launched the Vesuvius Challenge, a global AI competition offering $700,000 to anyone who could extract real text from the carbonized scrolls.
The winners?
Three students from three different countries who had never met in person — and who connected entirely online:
Luke Farritor, a 21-year-old undergraduate computer science student at the University of Nebraska–Lincoln, was the first to detect a Greek word in the scroll using a custom-trained model.
Youssef Nader, a PhD student in biorobotics at the Free University of Berlin, developed a different ink detection model that had earned second place in a previous prize round. After seeing Luke’s post about his model on Discord, he reached out to collaborate.
Julian Schilliger, a robotics student based in Zurich, Switzerland, had been working on segmentation and mapping tools independently and joined the team shortly afterward.
They met through the Vesuvius Challenge Discord server, initially as competitors. But after discovering their efforts were complementary, they began working together remotely. Their collaboration grew organically through GitHub, where they merged their models, code, and pipelines — coordinating across time zones and using tools like GitHub Copilot to accelerate development.
🔓 The Breakthrough
Their teamwork paid off. Using their combined models, they virtually unwrapped one of the scrolls and identified more than 2,000 Greek characters — including full words — from a scroll that hadn’t been read in nearly 2,000 years. Their submission made history, earning them the grand prize and global recognition.
🔍 What’s Next?
Hundreds more scrolls remain.
The next phase of the Vesuvius Challenge is live, with Seales calling the progress nothing short of a “renaissance.”
He believes the Villa of the Papyri, not yet fully excavated, may hold even more scrolls — possibly thousands.
“So much remains for us to discover,” says Seales.
🌍 Why It Matters That AI Reads Roman Scroll Without Touching It
This isn't just about technology — it’s about unlocking lost knowledge from the ancient world, potentially including unknown works of philosophy, science, and literature from the Roman era.
It’s also about the power of human connection in a digital age. Three students, working from Nebraska, Berlin, and Zurich — who had never met before — joined forces through curiosity, Discord, and GitHub. Together, they solved a problem that had stumped researchers for centuries.
It’s a powerful reminder:
When we combine ancient artifacts, cutting-edge AI, and collaborative minds, the past comes roaring back to life.
🧠 TL;DR:
1,800 ancient scrolls buried in Vesuvius eruption
AI + X-ray-like tech helped decode one without unrolling it
A student team from the U.S., Germany, and Switzerland cracked the first text in 2,000 years
They met through Discord and collaborated entirely online using GitHub
A new wave of discovery is just beginning
Enjoyed this article?
Stay ahead of the curve by subscribing to NewBits Digest, our weekly newsletter featuring curated AI stories, insights, and original content—from foundational concepts to the bleeding edge.
👉 Register or Login at newbits.ai to like, comment, and join the conversation.
Want to explore more?
AI Solutions Directory: Discover AI models, tools & platforms.
AI Ed: Learn through our podcast series, From Bits to Breakthroughs.
AI Hub: Engage across our community and social platforms.
Follow us for daily drops, videos, and updates:
And remember, “It’s all about the bits…especially the new bits.”
Comments