Last year, four researchers linked to Artificial Intelligence won the Nobel Prize, including the founder of DeepMind, Demis Hassabis, a company purchased by Google. AlphaFold, one of the company’s products, was able to predict three-dimensional structures of proteins, allowing scientists to understand, in a matter of months rather than decades, how they fold. The applications in the health sector alone are countless: discovering how a virus invades a cell, how a medicine finds its target, how a molecule defective triggers heart disease.
The engines at the heart of AI, neural networks, are not just answering questions, creating viral videos or funny images. In addition to the health sector, AlphaFold is being used to support many other applications, such as the design of new enzymes capable of degrading plastic in weeks (not centuries), a promise of a revolution in environmental management, or synthesizing proteins not found in nature, which function as small molecular machines, i.e. industrial catalysts clean, biological filters and environmental sensors.
Still regarding the environmental impact, studies are also advancing rapidly in beekeeping. Researchers use “machine” learning models to analyze bee genetic and behavioral patterns, with the aim of selecting colonies that are more resistant to pesticides, parasites and climate variations. The objective is to reduce losses and ensure the continuity of pollination, a critical factor for agricultural productivity.
In the electricity sector, AI makes networks more efficient by predicting consumption peaks and fluctuations in renewable generation in advance. Estimates combine historical data, weather and satellite imagery to calculate wind and solar panel production or demand in heat waves. In Brazil, operators already use these models to activate reserve plants or batteries before sudden changes, balancing the load and avoiding waste or the dreaded blackouts.
The startup Project Prometheus, led by Bezos, the billionaire founder of Amazon, aims to use AI to reinvent engineering and industrial production, from computers and automobiles to spacecraft. The company has already raised approximately 5800 million euros and brought together around 100 professionals from cutting-edge centers such as OpenAI, DeepMind and Meta. In this way, it seeks to create AI systems capable of guiding the entire manufacturing cycle: designing parts, optimizing assemblies, predicting failures and accelerating testswhich, in theory, can drastically reduce costs and development time.
A scientific and market revolution is multiplying our ability to understand, heal, produce and preserve. At the end of this decade, we may even have to work less and study more. It will be up to societies to adapt to these enormous changes and allow everyone to benefit from these advances.