This Scientist Uses AI to Hunt New Antibiotics — Even in Extinct Species
Drug-resistant germs kill over a million people a year. AI is hunting new weapons.
Antibiotics are quietly failing us. Bacteria are evolving ways to survive the drugs we've relied on for decades, and health experts warn that common infections could become dangerous again. The problem is that finding brand-new antibiotics is slow, expensive, and mostly done by hand in labs. That's where César de la Fuente, a researcher at the University of Pennsylvania, is trying something different: pointing artificial intelligence at the genetic code of living things — and of species that died out thousands of years ago.
The tools doing the heavy lifting are OpenAI's Codex, an AI that writes computer code, and ChatGPT. Think of Codex as a tireless programmer that builds the software needed to comb through enormous piles of genetic data, and ChatGPT as an analyst that helps read and reason about what turns up. Together they act like a research assistant that can scan millions of pages of biological information in hours and hand back a short list of molecules worth actually testing at a lab bench. That shortlist matters because lab testing is the slow, costly bottleneck in drug discovery.
Why look at extinct species? Every organism makes its own chemical defenses. A woolly mammoth or a Neanderthal carried molecules that modern bacteria have never encountered, so those germs haven't had a chance to build resistance. That makes ancient DNA a promising place to look. So far, the work has produced candidates that kill bacteria in the lab — encouraging, but very early.
The honest catch: none of this is a medicine yet. A molecule that kills bacteria in a dish still has to survive years of testing for safety and effectiveness in humans, and most candidates fail somewhere along the way. De la Fuente's approach doesn't replace that process; it tries to start it from a better place. If it works at scale, it could shorten the years it takes to find the next generation of antibiotics — the drugs that keep surgery, childbirth, and everyday infections from turning deadly.
- A researcher is using AI to scan the genetic code of living and extinct species for molecules that kill harmful bacteria
- Codex (an AI that writes code) handles the heavy data crunching; ChatGPT helps interpret the results, acting like a fast research assistant
- Ancient DNA from extinct species is attractive because modern bacteria have never seen those molecules, so they may not resist them
Why It Matters
New antibiotics could keep routine infections and surgeries safe as drug resistance grows — though human treatments are still years away.