Increasingly strong bacteria, but artificial intelligence can beat them
Growing data in Italy, but also new hope from bioinformatics and generative language models
Antibiotic resistance is one of the most serious health emergencies of the 21st century. Bacteria are becoming increasingly resistant to the drugs that used to neutralise them, and according to the European Centre for Disease Control (ECDC), Italy is among the countries most affected.
In 2024, there were approximately 12,000 deaths attributable to antibiotic-resistant infections, a number that places us at the top of the European league table. At the same time, the consumption of antibiotics in our country has grown: in 2023 it reached 22.4 defined daily doses per 1,000 inhabitants, an increase of 6.3 % compared to the previous year (Aifa data, 2025) and according to the European Agency's estimates 'has an annual cost for our SSN of EUR 2.4 billion, with 2.7 million beds occupied due to these infections,' says Aifa president Robert Nisticò.
Excessive and often inappropriate use contributes to a phenomenon that today no longer only affects hospitals, but also common infections of everyday life.
To tackle this threat, Italy has adopted the National Plan to Combat Antimicrobial Resistance (Pncar 2022-2025), based on the 'One Health' approach, which recognises the interconnection between human, animal and environmental health. However, preventive strategies must go hand in hand with innovation in the search for new antibiotics, an area that has suffered a major setback for decades.
And this is precisely where artificial intelligence comes in. Generative Ai models and machine learning are revolutionising the discovery of new molecules, reducing development time and costs. In several international projects, neural networks and language models are 'trained' on huge chemical databases to automatically identify potentially effective structures against multi-resistant strains.
