AI Reshapes Science and the Humanities, from AlphaFold to the Search for Historical Patterns

DeepMind’s AlphaFold cracked the protein-folding problem in 2020, allowing researchers to rapidly predict three-dimensional structures from amino acid sequences and shorten the search for drug targets. Meanwhile, “cliodynamics” uses statistical models to analyze wars, populations and the rise and fall of empires, extending AI’s reach from life sciences into historical research.

DeepMind unveiled AlphaFold 3 in May 2024, enabling predictions of interactions between proteins and DNA, RNA and drug molecules. At the time, the AlphaFold database contained more than 200 million protein structures. The work was recognized with the Nobel Prize in Chemistry on October 9, 2024, while generative AI was also being applied to the design of new proteins.