Beyond Shovels and Brushes: The AI-driven Evolution of Archaeological Exploration

It seems pretty clear right now that AI is likely the “next big thing.” This exciting (and perhaps scary) technology seems set to change the world and revolutionize our lives in a myriad of ways. But some of its potential applications may not be all that obvious at first glance. Archaeology is the study of human history through the excavation and analysis of artifacts and ancient structures. When most of us think of archaeology we think of people digging through dusty earth picking at bits of broken pot. How can AI help them? But by harnessing the power of AI, archaeologists are revolutionizing their field, enabling more accurate interpretations, accelerating discoveries, and enhancing our understanding of past civilizations. 

1.    Machine Learning for Genomic Dating

We start off with one of the most revolutionary uses for AI in archaeology. In the August of 2022, a scientific paper was published in Cell Reports Methods which laid out how a Lund University team from Sweden had used machine learning (AI) to identify genomes in dead organisms and work out how old they were.

Every living thing has a genome, a complete set of DNA, and in most creatures, it is present in nearly every cell. Being able to take a sample of a dead creature’s genome and use it to work out how old it is could be revolutionary, but why?

With AI, scientists will be able to precisely understand DNA data. (gerasimov174 / Adobe Stock)

Well, as the paper explained, accurate dating is vital if archaeologists are hoping to interpret paleo-genetic data correctly. Right now, we largely rely on radiocarbon dating, a method invented in 1946 by Willard Lobby. This method measures the amount of the isotope carbon-14 in an object to work out how old it is. The less carbon-14 it has, the older it is.

But radio-carbon dating has a major limitation, the process requires a high level of collagen extraction. This means, according to the researchers, around half of all published ancient genomes don’t have reliable and direct dates. This is obviously less than ideal. 

This is why the Lund team has created their temporal population structure (TPS), a machine learning-based genomic dating method for genomes. It can date anything ranging from the Late Mesolithic era to modern times with brilliant accuracy. 

But, the technology has one more benefit which will only grow with time. Currently, we can’t carbon date anything that was alive after 1950. This is because the testing of nuclear weapons between 1952 and 1962 released substantial amounts of carbon-14 into the environment and ever since living things have been soaking it all up. So, for future researchers, technology like TPS will be vital for dating remains.

• Inaccuracies Found in Radiocarbon Dating Calibrations Could Change Historical Timelines
• The Origin of Poop: AI to Predict Source of Ancient Feces

2.    Identifying Rock Art

AI has also been found to be incredibly useful for identifying and analyzing rock art. Ancient rock art is one of the most enigmatic areas of archaeology. It tends to be incredibly old, and we often have no idea what the strange symbols and pictures mean. There’s also a lot of it, which means there’s a lot of data to sift through.

Aboriginal rock art at Ubirr Art Site, Kakadu National Park, Northern Territory, Australia. (Luke Durkin/CC BY 2.0)

In a recent study based at the Kakadu National Park in Australia’s Northern Territory researchers used another machine learning model to detect art from hundreds of photos. Some showed actual rock art while others just showed bare rock faces.

The fledgling system managed to detect the art with an 89% success rate, which isn’t bad. The researchers hope that the technology will be useful for assessing large collections of images from heritage sites around the world. The possibilities are really exciting.

The nature of machine learning means that the more data the model is exposed to, the better it will become. In the future, it could be used to date newly found rock art by comparing it to what has been found before. Or it could be used to identify genuine ancient rock art against modern fakes or reproductions. Give it enough data and it could begin unraveling the mystery of the meanings behind rock art, in the same way, AI can be used to translate languages and decipher riddles (which we’ll come to later).

3.    Analyzing Satellite Imagery

We already covered that archaeologists spend a lot of time digging around in dusty earth but how do they decide where to dig? Why dig up this field and not that field? Well, locating potential excavation sites used to mean a lot of time-consuming fieldwork, visiting sites to see if they looked viable. But AI offers a more efficient option.

The first satellite images were beamed back to Earth in August 1959 from NASA’s Explorer 6 satellite. Ever since we’ve been using satellites to capture absolutely huge amounts of data, there’s not a square inch of the planet that hasn’t been photographed by now by multiple eyes in the sky. 

AI can be trained to rifle through and analyze this immense amount of data to identify potential archaeological sites. It can do this by discerning subtle patterns and anomalies in the imagery that may indicate the presence of buried structures or settlements.

The beauty of Earth from afar: an enchanting view generated by AI. (Viiiiiiiiiiiiiiiiiii/Adobe Stock)

So, from the comfort of their office or even home an archaeologist can sit behind their screen and potentially identify a promising area for a dig site. Better yet, the technology can be used to identify potential dig sites in remote, dangerous, or inaccessible areas. 

Imagine, for example, an archaeological team is interested in an ancient civilization that lived in the heart of the rainforest or desert. Traditionally they’ve had to organize an expensive (which means seeking funding) and time-consuming expedition just so that they could do their fieldwork. Fieldwork during which they might come up empty-handed. AI analysis of satellite imagery can help cut down the risk, giving them a good idea of where to look.

It can also be used to provide extra information on the landscape surrounding a site. This extra data can help historians and archaeologists understand the context in which societies lived, potentially revealing more about their cultures and lifestyles.

4.    AI and Archaeology: Going Through Geological Data

And it’s not just satellite images AI can be used to go through. We also have immense amounts of geological data. Archaeologists can use this to analyze the different layers of our planet’s crust for evidence of past human activities.

For example, soil samples can show what crops were planted in a specific area, highlighting the agricultural practices of ancient civilizations. Or stratigraphy, looking at the geological layers of strata and sediments can help archaeologists establish the chronological sequence of different archaeological sites. By analyzing the types of sediment and the materials within each layer, archaeologists can determine the relative age of artifacts and structures found in different layers. This helps create a timeline of human activity.

Chalk layers in Cyprus - showing classic layered structure. With AI archaeologists can easily identify the age of structures and artifacts. (Public Domain)

But this just scratches the surface, geological data can also be used to: work out why people settled where they did, identify geological processes like erosion and tectonic events, identify ancient trade networks, identify natural resources and how our ancestors exploited them, and reconstruct ancient landscapes. Safe to say, archaeologists rely on it quite a lot.

The problem is there’s a lot of this data to get through and it can be incredibly complex, especially in areas that have seen thousands of years of human activity. There are also high degrees of variation from place to place. This is where AI comes in.

Much like satellite images, machine learning can be used to analyze patterns and anomalies in our collected data. It’s particularly good at spotting things humans might have missed the first time around. On the one hand, this means it can simply be used to make archaeologists' lives easier, helping cut down the amount of time they spend looking at reams of geological data. On the other hand, it could be used to find previously unknown sites.

5.    Preservation

AI isn’t just useful for making new discoveries. One of the biggest challenges archaeologists often face is protecting what has already been discovered. Ancient sites face various threats, both natural and manmade.

AI algorithms can be used to monitor changes in the landscapes surrounding archaeological sites. They can then identify potential threats like illegal excavations or climate change impacts. Rising sea levels, increased erosion (like from acid rain), and urban development can all threaten important sites. AI can provide an early warning system for all of these things so that preventative measures can be put in place before the damage is done. 

And this technology is already being used. Baiae is a sunken Roman city around 50 miles away from Pompeii. It’s now home to a network of acoustic modems and underwater wireless sensors that are being used to capture environmental data and transmit it to land in real time where machine learning analyzes it. 

Everything from actual pictures to information on water quality, pressure, and temperature to noise levels, water currents, and waves is measured. The AI can then monitor and look out for damage caused either by divers or the above environmental factors. 

Not impressive enough? Well, Pompeii, one of the most famous archaeological sites in the world, is currently being guarded by an AI-powered robot dog, Spot, built by Boston Dynamics. He (or she) is the ultimate guard dog. Spot and his AI are being used to wander around Pompeii and look for any structural/ safety issues as well as being sent underground to sniff out illegal relic hunters.

The quadruped robot from Google's Boston Dynamics. (Steve Jurvetson/CC BY 2.0)

6.    AI and Underwater Robots

But it isn’t just in preserving underwater sites that AI can lend a hand, it can help us explore them too. If the 2023 Titan submarine incident taught us anything it’s that there are still inherent risks to exploring underwater areas.

There’s a surprising amount of human history submerged beneath the waves and archaeologists are starting to use autonomous underwater vehicles (AUVs) to chart and uncover underwater archaeological sites in high detail. AI is great at interpreting sonar data, detecting dangerous anomalies, and finding things like shipwrecks and underwater cities. 

7.    Language Decipherment and Translation

Inscriptions and ancient texts often provide invaluable insights into the cultures and societies of the past. The problem is that there are numerous ancient languages like Linear A, Rongorongo, and Etruscan that we either haven’t been able to decipher at all or don’t have a complete understanding of. 

A large part of deciphering an ancient language is pattern recognition, which AI excels at. Today, AI-driven language decipherment tools are assisting archaeologists in deciphering ancient scripts and languages by identifying patterns, characters, and linguistic relationships. This breakthrough technology is unlocking lost knowledge from civilizations that left behind written records.

Detail of an unidentified plant on page 78r of the Voynich Manuscript. The hand-written codex remains undeciphered.  (Public Domain)

Besides language translation, it’s hoped that AI will help us decode texts like the Voynich Manuscript, a medieval document containing mysterious text and illustrations. It has puzzled linguists and cryptographers for centuries, and its contents have yet to be definitively deciphered, maybe AI will be able to do what humans can’t. 

At its simplest level, AI is also used for writing reports, academic papers, essays, and even grant applications, generating in seconds what would ordinarily take students and scholars days, weeks, or even months. Traditionally, writing services have also done this task, but many websites are not trustworthy, so it is important to seek out reviews to the most reliable sites, for instance, the website www.writingpapersucks.com/papersowl-com-review/

8.    Virtual Reconstruction

AI gets a lot of negative press and much of this is in relation to “deep fakes” and AI-generated art. Basically, give an AI enough data and clear instructions and it can create some pretty impressive and incredibly detailed images. Images realistic enough to trick the human eye.

AI-powered virtual reconstruction techniques are breathing life into ancient sites that have deteriorated over time. By analyzing archaeological data, historical records, and artistic representations, AI algorithms can create immersive 3D visualizations of how these sites appeared in their prime, enabling researchers and the public to explore ancient worlds in a digital environment.

AI in archaeology: dozens of images layered to create this virtual reality replica of a 16th century church in Peru. (Vanderbilt University)

It’s not quite a time machine, but it’s the next best thing. Combine these models with VR (Virtual Reality) technology and we have an incredibly exciting gateway to the past. Archaeologists can have a hard time getting the public excited about their discoveries, this could be the way to do it. 

• Google AI Decodes Broken Greek Texts Better than Humans
• High-Technology Discovered in Classical Mythology Reveals The Ancient Origins of Artificial Intelligence

Conclusion

It looks like AI is set to change a lot of professions of future and archaeology is one of them. It’s likely that in the future archaeologists will spend much of their time behind a screen acting as data scientists, training neural networks to help them find new dig sites and explain what they’ve already discovered.

Which might seem a little dull to some of us. After all, Indiana Jones carries a whip, not a laptop. But fear not. The role of an archaeologist might be evolving but at the end of the day, someone is always going to need to get their hands dirty and do some actual fieldwork. That’s something an AI can’t do.

In the right hands, AI is a powerful tool. For archaeologists, it can speed up the process of discovery and help protect what has already been found. By marrying AI and Archaeology, past and future, who knows what we’ll discover? These are exciting times indeed.

* This article contains sponsored links.

Top image: AI robot as an archaeologist, uncovering an ancient artifact.  Source: Hatia/Adobe Stock        

By Robbie Mitchell