As a teenager, I wanted to become an archaeologist and decipher ancient inscriptions. Many years, many trips, and many museum visits later, I know how difficult it is to assign meaning to letters, symbols, and words on weathered stone tablets. No matter how much progress researchers have made, many gaps remain, making it difficult to understand past times. Fragments are sometimes all that remain to bear witness to past eras and the fates of those who lived in them.
What if we could build a bridge to this distant past? What if state-of-the-art technology could help us read silent testimonies and bring lost stories back to life?
This is exactly what is happening at the intersection of archaeology and artificial intelligence. Aeneas, a new AI model from Google DeepMind, is turning this thought experiment into reality.
Aeneas deciphers texts
How much do you know about ancient Greek mythology? I had to look up who Aeneas was: a Trojan prince who escaped from the burning city with other Trojans, fled westward, and, according to Roman tradition, was one of the ancestors of Romulus and Remus, the mythological founders of the city of Rome. Now a new chapter is beginning for the legendary hero.
Google DeepMind's Aeneas project is an example of how modern technology is providing new tools for historical research. Developed in close collaboration with historians, this AI is designed to decipher ancient Greek inscriptions and thus expand our understanding of antiquity.
This AI model has two key capabilities:
Restoring texts: Aeneas can reconstruct missing letters or entire words in damaged inscriptions with impressive accuracy. The model draws on its extensive training with ancient Latin and Greek texts and quickly recognizes patterns that researchers may overlook.
Chronological classification: The AI can determine the date of an inscription's creation – often more accurately than conventional dating methods. This ability is particularly valuable when other clues for dating are missing or contradictory.
Humans and machines working in tandem
Despite all the enthusiasm about the technological possibilities, one thing remains crucial in a historical context: Aeneas does not replace human expertise. Aeneas is a powerful tool that supports and accelerates the work of historians and archaeologists, making it a suitable example of successful collaboration between humans and machines.
AI takes care of the detailed work of initial analysis and reconstruction, which is very time-consuming for human actors. This gives researchers more space for what machines cannot (yet) do: interpretation of content, historical classification, and a deeper understanding of cultural contexts.
Why connecting the old and the new is important
Every restored inscription is more than just an academic exercise. It opens a window to the past and gives us new insights into ancient times. When Aeneas reconstructs a missing text, it opens up access to forgotten worlds: we can learn how ancient societies formulated their laws, what trade relations they maintained, which gods they worshipped, and what everyday life was like for ordinary people. The fragments tell stories of hopes, conflicts, and achievements that would otherwise be lost forever.
The true power lies in the entirety of the inscriptions. The thousands of texts used to train Aeneas form a vast historical data set, a kind of collective memory of antiquity. Individual inscriptions may be only fragments, but when taken together, they reveal patterns and connections that would remain hidden to the human eye with the help of artificial intelligence. Linguistic developments, trade routes, social upheavals – all of this becomes visible when AI sifts through mountains of data and establishes connections.
Why do we need to understand the past in order to understand the present? The past is a mirror in which we can view our present. When we understand how past societies were organized, how they functioned, and why they failed, we gain new perspectives on our own challenges. Ancient texts show us with astonishing clarity that many of our current debates about democracy and civil rights, migration and integration, war and peace have a long history. Even the ancient Greeks discussed issues that preoccupy parliaments and the media today.
Lessons for the future?
Historical data could reveal recurring patterns of human behavior that have remained remarkably consistent over thousands of years. We could see how societies respond to crises: with innovation or retreat, with solidarity or division. We could see under what conditions people work together and when they come into conflict. These patterns do not necessarily repeat themselves, but they can provide us with valuable guidance. Or as Mark Twain is said to have put it: History may not repeat itself exactly, but it does rhyme.
Furthermore, knowledge of a shared, deep past can connect cultures. Recognizing that we all stand on the shoulders of previous generations and face similar challenges can, at best, foster empathy and a sense of global community. The ancient Greeks and Romans are not “the others”—they are part of our shared human history, whose challenges continue to shape us.
Perhaps the most important lesson from this project is the importance of knowledge preservation and collaboration. Just as Aeneas preserves knowledge from oblivion, we must work today to secure our collective memory and collaborate across disciplines and cultures—just as AI researchers and historians are doing in this project.
