Antibiotic resistance is no longer a future threat: it is a current crisis that causes more than a million deaths a year and complicates millions of additional treatments worldwide.
The problem has two main drivers: the overuse of antibiotics and the lack of new drugs.
When bacteria survive insufficient doses, they evolve defense mechanisms that render previously life-saving medicines ineffective.
Slow Diagnosis, Blind DecisionsOne of the biggest risks is time. Traditional methods for identifying resistant infections can take between 2 and 3 days. In critical diseases like sepsis, every hour of delay significantly increases the probability of death.
Given this delay, doctors often act by approximation, which increases the incorrect use of antibiotics and accelerates resistance even further.
AI as Structural ChangeAccording to Ara Darzi, the healthcare system is entering an inflection point.
New AI-based tools are achieving:
- Diagnostics with over 99% accuracy without complex infrastructure
- Rapid identification of resistant bacteria
- Large-scale spread pattern analysis
In one specific case, an AI system managed to decipher resistance mechanisms in 48 hours, something that previously took years of research.
Impact on Drug DevelopmentAI is also accelerating the discovery of antibiotics:
- Screening billions of molecules in days
- Designing completely new compounds using generative models
- Automating laboratory experiments
This allows for a drastic reduction in development times, one of the main historical bottlenecks.
The Problem is Not Technological, It Is EconomicDespite technical advances, pharmaceutical companies are abandoning antibiotic development. The reason is structural: these drugs must be used sparingly to prevent resistance, making them unprofitable.
To compensate, some countries are testing new models: