Archaeologists are turning to artificial intelligence to decipher a 2,000-year-old Roman inscription, and what’s written on the stone could change everything we thought we knew about an ancient lost game

The article discusses a limestone slab from a Dutch museum, known as “Object 04433”, which was previously unidentified.
Researchers concluded that the artifact was most likely a Roman game board, using microscopic wear analysis and artificial intelligence simulations.
The unique shape of four diagonals and one straight line on the stone’s surface prompted much debate, but wear patterns indicate it was used for a game.
An AI system called Ludii helped the researchers simulate game rules based on the stone’s wear, leading to the discovery that it was likely a game board used in a blocking strategy.
The artifact suggests that the tradition of blocking games existed in Europe during the Roman era, even earlier than previously documented.
The study implicates the repurposing of architectural stone, commonly used in the Roman era, indicating that the artifact might be a reused decorative piece.
Researchers believe that non-invasive methods like imaging and simulations can aid in identifying other historical objects, contributing to digital conservation efforts.
While AI is a powerful tool, it relies on human-defined rules and acknowledges that player variations can affect interpretations of wear.