Charred, the hundreds of fragile ancient parchments would fall apart if anyone tried to unroll them, and any trace of writing would be virtually illegible. The scrolls of Herculaneum, as they are known, still remain unopened, but thanks to the powerful tool that is artificial intelligence, their contents are now within reach.
Using AI and high-resolution X-rays, — the remarkable feat revealed the first complete passages of papyri that survived the eruption of Mount Vesuvius in 79 AD
The artifacts, recovered from a building believed to have been the home of Julius Caesar’s father-in-law, form an unprecedented collection of information about ancient Rome and Greece.
Computer scientists who launched the Vesuvius Challenge, a competition designed to speed up the decryption process, hope that 90% of four scrolls will be unlocked by the end of 2024. The main challenge has been to virtually flatten the documents and distinguish the black ink from the charred papyri to make Greek and Latin script legible.
“AI is helping us amplify the legibility of ink evidence,” said Brent Seales, a computer science professor at the University of Kentucky who has been working to decode the scrolls for more than a decade. “The paint evidence is there. It’s buried and camouflaged in all this complexity that AI distills and condenses.”
The project is a compelling example of the growing utility of artificial intelligence, which was crowned in 2024 with the Nobel committee recognizing the development and application of AI in science for the first time: , paving the way for how artificial intelligence is used today.
A fuzzy and often overrated term, AI aims to mimic human cognitive functions to solve problems and complete tasks. Artificial intelligence encompasses a range of computational techniques: using datasets to train and improve machine learning algorithms and enabling them to identify patterns and report predictions.
Some AI tools can pose risks, such as systems used in hiring, policing and loan applications that replicate biases, as they can be trained on historical data that reflects biases, for example about sex or race, that result in discrimination.
AI has transformed the landscape of scientific discovery, with the number of peer-reviewed articles using AI tools increasing dramatically since 2015, and those that use AI methods are more likely to be among the most cited. More than half expected AI tools to be “very important” or “essential” to research practice.
However, the world’s oldest science academy has warned that the black-box nature of many AI tools is limiting the reproducibility of AI-based research. For Seales, however, it is a powerful instrument deployed wisely that has yielded dramatic results.
“AI is a field of computer science designed to try to solve problems in ways that we thought only humans could solve,” Seales said. “I think of the kind of AI we’re using as a kind of superpower that allows us to see things in data that human eyes wouldn’t be able to see.”
The Vesuvius Challenge is just one way the rapidly evolving field has shaken up science and revealed the unexpected in 2024. AI is also advancing scientists’ understanding of how animals communicate in the deep ocean, helping archaeologists find new sites in remote and inhospitable terrains, and solving some of biology’s greatest challenges.
Decoding “Whale Language” and Other Animal Languages
Researchers know that the cryptic clicks made by sperm whales vary in timing, rhythm and length, but what the animals are saying with these sounds—produced through spermaceti organs in their bulbous heads—remains a mystery to human ears.
Machine learning, however, calls codas, which represent the voices of approximately 60 sperm whales in the Caribbean Sea. The work may one day make it possible for humans to communicate with marine animals.
Scientists have examined the timing and frequency of codas in solitary whale vocalizations, in choruses, and in call-and-response exchanges between the marine giants. When visualized with artificial intelligence, previously unseen coda patterns emerged in what researchers described as similar to phonetics in human communication.
In total, the program detected 18 types of rhythm (the sequence of intervals between clicks), five types of tempo (the duration of the entire coda), three types of rubato (variations in duration), and two types of ornamentation — an “extra click” ” added to the end of a coda in a group of shorter codas.
These features could all be mixed and matched to form a “huge repertoire” of phrases, scientists reported in May. However, the approach has its limitations. While machine learning is adept at identifying patterns, it does not clarify meaning.
A next step, according to the study, is interactive experimentation with whales, along with observations of whale behavior, which could be an important part of unraveling the syntax of sperm whale click sequences.
The approach could also be applied to other animals’ vocalizations, Dr. Brenda McCowan, a professor at the University of California Davis School of Veterinary Medicine, previously told CNN. She was not involved in the study.
Finding archaeological sites
Meanwhile, on land, artificial intelligence is now turbocharging the search for mysterious lines and symbols etched into the dusty soil of Peru’s Nazca Desert, which archaeologists have spent nearly a century discovering and documenting.
Often visible only from above, the sprawling pictograms depict geometric designs, humanoid figures, and even a knife-wielding orca.
A group of researchers led by Masato Sakai, professor of archeology at Yamagata University in Japan, trained an object detection AI model with high-resolution images of the 430 Nazca symbols mapped as of 2020. The team included researchers from the Center for IBM’s Thomas J. Watson Research in Yorktown Heights, New York.
Between September 2022 and February 2023, the team tested the accuracy of their model in the Nazca Desert, surveying promising sites on foot and using drones. The researchers ultimately “spot-checked” 303 figurative geoglyphs, nearly doubling the known number of geoglyphs in a matter of months.
The model was far from perfect. He suggested a staggering 47,000 potential sites in the desert region, which covers 629 square kilometers. A team of archaeologists filtered and classified these suggestions, identifying 1,309 candidate sites with “high potential.” For every 36 suggestions made by the AI model, the researchers identified “one promising candidate,” according to the study.
Nevertheless, AI has the potential to make enormous contributions to archaeology, particularly in remote and hostile terrains like deserts, even if the models are not yet completely accurate, said Amina Jambajantsan, a researcher and data scientist in the Institute’s archeology department. Max Planck of Geoanthropology in Jena, Germany.
Jambajantsan was not involved in the Nazca research, but uses an AI model to identify burial mounds in Mongolia based on satellite images.
“The problem is that archaeologists don’t know how to build a machine learning model and data scientists typically aren’t really interested in archeology because they can make a lot more money elsewhere,” Jambajantsan added.
Understanding the Building Blocks of Life
AI models are also helping researchers understand life on the smallest scale: chains of molecules that form proteins, the building blocks of life.
Although proteins are built from only about 20 amino acids, these can be combined in almost infinite ways, folding into highly complex patterns in three-dimensional space. The substances help form hair, skin and tissue cells; they read, copy and repair DNA; and help transport oxygen in the blood.
For decades, decoding these 3D structures has been a challenging and time-consuming endeavor involving the use of complex laboratory experiments and a technique known as X-ray crystallography.
However, in 2018, a revolutionary AI-based tool emerged onto the scene. The latest version of the AlphaFold Protein Structure Database, developed by Demis Hassabis and John Jumper at Google DeepMind in London, predicts the structure of almost all 200 million known proteins from amino acid sequences.
Trained with all known amino acid sequences and experimentally determined protein structures, the database works like a “Google search”. It provides one-button access to predicted protein models, accelerating progress in fundamental biology and other related fields, including medicine. The tool has already been used by at least 2 million researchers around the world.
“It really is an independent revolutionary breakthrough that solves a traditional holy grail of physical chemistry,” Anna Wedell, a professor of medical genetics at the Karolinska Institute in Sweden and a member of the Royal Swedish Academy of Sciences, told CNN after Hassabis and Jumper were among the three winners of the 2024 Nobel Prize in Chemistry.
The tool has some limitations. Attempts to apply AlphaFold to proteins based on mutated sequences, including one linked to early breast cancer, have confirmed that the software is not equipped to predict the consequences of new mutations in proteins.
AlphaFold is just the best-known AI tool among several being deployed in biomedical fields. Machine learning is accelerating efforts to compile an atlas of every cell type in the human body and discovering molecules that become new drugs, including a type of antibiotic that could work against a particularly threatening drug-resistant bacteria.
*Mindy Weisberger and Taylor Nicioli contributed to this report
