AI unseals charred 2,000-year-old Vesuvius scrolls
A trio of student researchers just used AI to read over 2,000 Greek letters inside a carbonized scroll from Mount Vesuvius's AD79 eruption, winning the $700,000 Vesuvius Challenge grand prize. The team applied machine learning to high-resolution CT scans to virtually unroll the fragile papyrus and detect ink traces. It opens the door to reading hundreds of completely charred ancient texts once thought lost forever.

For over two centuries, scholars have stared helplessly at hundreds of lump-like, charred papyrus rolls pulled from a Roman villa buried by Mount Vesuvius in AD79. Touch them, and they crumble into ash. Try to read them, and you find black carbon ink burned onto blackened paper.
That ancient roadblock just fell. Nat Friedman and the organizers of the $1m Vesuvius Challenge announced that a three-person student team has won the $700,000 grand prize after using artificial intelligence to read more than 2,000 Greek letters inside an unopened, carbonized scroll.
High-energy physics meets student hackers
The breakthrough rests on work led by computer scientist Brent Seales at the University of Kentucky. Seales spent years building techniques to virtually unwrap the fragile layers using high-resolution CT scans captured at the Diamond Light Source, the UK’s national synchrotron. Crucially, his team figured out that AI could detect microscopic textural changes left by ink on the papyrus fibers.
Instead of keeping the research behind university walls, Seales teamed up with tech investors to launch an open prize. When a storm nearly grounded the scroll’s flight to the particle accelerator, Friedman chartered a private jet to save the slot.
Global teams answered the call. In October, US student Luke Farritor claimed early prize money by detecting the word for "purple". He then teamed up with Youssef Nader in Germany, and Julian Schilliger in Switzerland joined right before the December 31 deadline with an algorithm that automated the virtual unwrapping. Working around the clock, the trio unlocked full columns of text.
Papyrologists were floored. Bristol University professor Robert Fowler believes the author is likely the Epicurean philosopher Philodemus. The text muses on the nature of pleasure, food, music, and specifically capers, debating whether joy in a mix of elements comes from the abundant parts or the scarce ones, before taking parting shots at philosophical rivals.
Why it matters
We think this is one of the most consequential demonstrations of machine learning outside pure enterprise software.
For historians and humanities researchers, it unlocks a literal time capsule. Herculaneum holds the only intact library to survive from antiquity, with hundreds of rolls already excavated and potentially thousands more still underground. Beyond this villa, museums worldwide hold crates of unreadable papyri, including sheets salvaged from Egyptian mummy cartonnage containing everyday tax receipts and letters. AI has just turned dusty coal back into historical records without destroying a single fiber.
It also highlights a brilliant blueprint for scientific progress. Combining national lab-grade scanning hardware with open-source base algorithms and Silicon Valley bounty incentives mobilized distributed talent that academic grants rarely reach. Three students outpaced traditional departmental timelines in under a year.
The catch? We have only read roughly 5% of this single scroll so far. The 2024 goal is to reach 85%, and scaling this pipeline to hundreds of damaged rolls will demand fully automated surface segmentation and even sharper ink-detection models. Still, antiquity's largest lost library is finally coming back online.
Sources (1)
- theguardian.com(opens in a new tab)Originaltheguardian.com