Empowering Science through Technology: Three AI-Driven Discoveries of the Year

In 2024, artificial intelligence has made it possible to advance science thanks to discoveries that only it could make possible, all synthesized by CNN.

Language decoding in sperm whales

For many years, researchers have been trying to identify what the “clicks” are, sounds produced by sperm whales, produced by the spermaceti organs located in the skull. Used to find your way underwater and hunt. But thanks to AI, we know that these sounds are also used for something else.

Indeed, after analyzing 9000 codas (sequences of clicks), scientists, using this tool, were able to realize that these sounds were also used by these mammals to communicate, with a number of different rhythms (18 ), or even different tempos (5). Although the content of these exchanges cannot yet be understood, knowing that sperm whales use clicks to communicate is a big step forward. Which could why not in the future allow man to communicate with marine animals.

Deciphering parchments almost 2000 years old

How many ancient texts have been found without ever being able to be deciphered? Whether it is because the paper is too damaged and crumbles at the slightest contact, or because the writing has not held up over time, many specialists have not been able to properly study these traces of history. Well that won't really be a problem with AI anymore. Indeed, researchers are on the verge of succeeding in deciphering four scrolls dating from 79 AD.

The scrolls of Herculaneum, almost burned after the eruption of Vesuvius, can now be exploited thanks to AI and high-resolution X-rays. Brent Seales, a computer science professor at the University of Kentucky, told CNN that “the evidence for the ink is there. They are buried and camouflaged in all this complexity that AI distills and condenses. Researchers hope that 90% of these four scrolls, recovered from what is believed to have been the house of Julius Caesar's father-in-law, will be deciphered before 2025.

AI used to understand proteins

Last example and not least, the study of proteins. These biological macromolecules are present in all living cells and are essential for the proper functioning of our body, such as building bones or skin, or even repairing DNA and oxygenating our blood. But their structure and size make them difficult to decode.

But since 2018, an AI-based tool has made it possible to have a database for proteins. This year, the researchers who developed AlphaFold made an improvement to this artificial intelligence tool. The latter can predict the structure of almost all of the 200 million known proteins from amino acid sequences. A breakthrough for the scientific world, which earned them the 2024 Nobel Prize in Chemistry.

Source: CNN