AI and History Unite: How Aeneas is Transforming the Study of Ancient Roman Inscriptions

AI Meets History

In the world of artificial intelligence, we often hear about breakthroughs in language models, image generation, or robotics. But sometimes, AI takes a quieter yet equally meaningful turn—like helping us better understand our shared human history. That’s exactly what’s happening with Aeneas, a new AI model developed to assist historians in interpreting ancient Roman inscriptions. Created through a collaboration between researchers at DeepMind and several universities across Europe, Aeneas is designed to support historical research by making sense of fragmented and weathered texts from the Roman Empire. For those who’ve ever marveled at how we piece together stories from the past, this development offers a fascinating glimpse into how technology can enhance our understanding of history.

The Name Behind Aeneas

Aeneas is named after the legendary hero of Roman mythology, and fittingly so—it navigates complex terrain to uncover meaning from scattered remnants of the past. At its core, Aeneas is an AI model trained on over 176,000 Latin inscriptions collected from digital epigraphic databases. These inscriptions range from official decrees and tombstones to everyday graffiti and personal letters. Many are incomplete or damaged by time, making them difficult to interpret without expert knowledge and context.

Innovative Analysis Techniques

What sets Aeneas apart is its ability to analyze both text and images of inscriptions simultaneously—a technique known as multimodal input. This allows it not only to restore missing parts of a text but also to estimate when and where it was created. One particularly useful feature is its “parallels search,” which finds similar inscriptions based on language patterns, structure, and historical context. This helps historians quickly identify connections that might otherwise take months or years to uncover.

Understanding Limitations

Of course, no tool is perfect. While Aeneas performs impressively—restoring gaps in texts with up to 73% accuracy when the missing section is ten characters or fewer—it still has limitations when dealing with highly ambiguous or severely damaged material. Its predictions are probabilistic rather than definitive, meaning they offer likely interpretations rather than final answers. However, many historians see this as a strength rather than a weakness; it encourages collaborative analysis rather than replacing human judgment.

Building on Previous Work

To appreciate how far this project has come, it’s helpful to look back at its predecessor: Ithaca. Released earlier by the same team, Ithaca focused on ancient Greek inscriptions and laid the groundwork for using AI in historical restoration tasks. Aeneas builds on that foundation by expanding capabilities into Latin texts and adding contextual interpretation features that go beyond simple restoration. The shift from Ithaca to Aeneas reflects a broader trend in AI development—moving from single-task tools toward more integrated systems that can support complex human workflows.

Open Access for All

This latest release also aligns with growing interest in applying generative AI beyond commercial or entertainment uses. In recent years, we’ve seen similar initiatives aimed at digitizing cultural heritage or supporting academic research through machine learning tools. What makes Aeneas stand out is its open-access approach: both the model and its dataset are freely available for researchers, educators, and museum professionals through an interactive platform called predictingthepast.com.

A Collaborative Future

In closing, Aeneas represents a thoughtful application of AI—one that respects the complexity of historical research while offering practical tools to support it. Rather than replacing historians, it works alongside them as a kind of digital assistant that brings speed and scale to tasks traditionally done by hand. For anyone interested in how technology can help us better understand where we come from, this project offers an encouraging example of innovation rooted in collaboration and curiosity. As AI continues to evolve, efforts like Aeneas remind us that some of its most meaningful contributions may lie not just in shaping our future—but also in illuminating our past.