Exploring History with AI: The Opportunities and Challenges of Digital Time Travel

Will AI Enhance Our Understanding of Human History or Signal Its End?

As we explore the intersection of artificial intelligence (AI) and history, compelling evidence emerges to support both sides of the debate.

AI Unlocks the Past

Recently, three computer science students won a prestigious award for their innovative use of AI, specifically machine learning and neural networks, to decipher an ancient Roman scroll from Herculaneum, near Pompeii.

These scrolls, part of a larger collection recovered from a villa believed to have belonged to Julius Caesar's father-in-law, represent the largest surviving library from classical antiquity. Between 1752 and 1754, archaeologists uncovered around 1,800 fragments of papyrus scrolls, possibly corresponding to 1,000 original volumes. Damaged during the eruption of Mount Vesuvius in A.D. 79, the scrolls have been compared to charred logs from a campfire.

According to the Herculaneum Society, these fragile, carbonized relics may contain lost philosophical dialogues and insights capable of reshaping our understanding of the ancient world. Previous attempts to read the scrolls have failed due to their delicate condition.

The breakthrough came from contestants in the Vesuvius Challenge, initiated in March 2023, where global participants raced to interpret the ancient texts. The winners virtually unspooled the scroll and successfully read over 2,000 Greek letters. Reports suggest the text is a previously unknown work by Philodemus, an Epicurean philosopher from Herculaneum, raising hopes that other scrolls could soon be deciphered.

Robert Fowler, a leading classicist and chair of the Herculaneum Society, noted that these texts could dramatically alter our understanding of key historical periods. This discovery exemplifies how AI is not only unlocking the past but deepening our knowledge of it. In addition to deciphering scrolls, deep learning techniques are being applied in archaeology to classify pottery fragments and locate shipwrecks using sonar imaging. One notable initiative, the “RePAIR” project, aims to reconstruct ancient artworks in Pompeii, starting with two-thousand-year-old frescoes.

AI: A Threat to History?

While these applications of AI in historical research are promising, there are concerns about its broader implications for human culture and history. Yuval Noah Harari, historian and author of Sapiens, warned that AI could absorb all human culture and potentially herald the end of an era dominated by human beings. This presents a dual-edged nature of AI, capable of both enriching and fundamentally changing our cultural landscape.

Harari posits that AI could reconstruct history through its own analyses, overriding established interpretations and potentially reshaping our understanding of the past. This raises the prospect of AI becoming a dominant force in societal narratives, compelling humanity to adapt to its influence.

The Need for Digital Integrity

To safeguard the integrity of our historical record, better preservation practices are essential. As demonstrated by the Herculaneum scrolls, the medium used for storage plays a critical role in preservation success. Merely storing documents digitally is inadequate, as digital formats can also degrade over time.

One promising solution is Microsoft’s Project Silica, which uses glass for data storage, potentially ensuring data longevity for thousands of years. This could serve as an effective method for archiving historical records prior to the advent of AI.

However, not only does the threat of an alien intelligence loom, but human manipulation of historical narratives is also a pressing concern. A recent New York Times op-ed highlighted the risks posed by generative AI, which can fabricate past events just as easily as it can fake current ones. The authors caution that false documents can serve as tools for distorting the historical narrative.

For instance, researchers at MIT created a deepfake of Richard Nixon delivering a speech prepared in the event of a disastrous moon landing, illustrating how technology can distort our collective memory. Francesca Panetta from MIT remarked that this “alternative history” showcases the capability of new technologies to obscure the truth, urging audiences to critically evaluate the media they encounter.

The potential for deepfakes to undermine trust in both present-day and historical contexts emphasizes the urgent need for strict guidelines regulating AI-generated content.

If we wish to steer the trajectory of AI and its impact on history, action must be taken swiftly. Present choices will shape future developments. It is crucial to navigate these advancements responsibly, ensuring AI serves as a catalyst for enlightenment rather than confusion and deception.

Rather than marking the end of human history, AI could serve as a springboard for deeper understanding and insight.

Gary Grossman

EVP of Technology Practice, Edelman

Global Lead, Edelman AI Center of Excellence

DataDecisionMakers

A hub for experts to share innovations and insights related to data and technology. Join us for cutting-edge ideas and best practices shaping the future of data. Consider contributing your own article!

Most people like

Find AI tools in YBX