AI SCIENCE
Reverse-engineering Roman entertainment with AI
AI has made its way into almost every industry, but archaeology has mostly stayed hands-on.
Now, new research shows it could play a role there too.
A study published in Antiquity used AI to help figure out how a Roman board game may have been played.
We already know board games existed in the ancient world through texts, artwork, dice, and playing pieces.
But actual game boards are harder to understand.
They are often found without pieces or instructions, which makes their purpose unclear.
That is what happened with a carved limestone object found in the Roman town of Coriovallum, now Heerlen in the Netherlands.
It had been seen as a possible game board for years, but no one knew for sure.
So researchers studied the wear on its surface and paired that with AI simulations.
In brief:
The stone showed wear patterns that suggest repeated game play.
AI simulations helped narrow down the most likely rules.
The best match was a Roman blocking game.
Rome.exe loaded
They found that the carved lines were more worn than the rest of the stone, suggesting that hard objects had been moved across it many times.
That gave weight to the idea that it had been used for a game.
The team then used an AI-supported system called Ludii to test possible rule sets.
After several simulations, they found that the closest match was a blocking game where one player tries to trap the other.
The research shows AI could help archaeologists do more than identify ancient objects. It may also help explain how people once used them.
We really had to ask AI what the Romans were doing with their weird little stone slab.- MG