Could artificial intelligence be the hero in our fight against superbugs?
Researchers at MIT think so. They’ve just unveiled two brand-new antibiotic candidates — designed by AI, atom-by-atom.
They wiped out drug-resistant gonorrhoea and MRSA in lab and animal tests.
For decades, antibiotic discovery has stalled while bacteria have evolved deadly resistance, causing over a million deaths each year.
“AI can enable us to come up with molecules cheaply and quickly.”
“And give us a leg up in the battle of our wits against superbugs,” says Prof James Collins of MIT.
What Exactly Did They Do?
The team trained their generative AI on chemical structures and bacteria-growth data, then let it create — not just search for — new molecules.
It scanned 36 million possibilities, filtered out anything too toxic or too similar to existing drugs, and delivered two promising designs.
But here’s the catch: they’re years away from pharmacy shelves.
Refinement, safety checks, and costly clinical trials still stand in the way.
“We still need to do the hard yards,” warns Dr Andrew Edwards of Imperial College London.
And even if they work, there’s a paradox: the best way to preserve a new antibiotic’s power is to use it sparingly — which makes it hard to profit from.
So, can AI save us from superbugs? Yes. Will it be easy? Not a chance.
