Generative AI boost in battle against drug resistance
15 Aug 2025
Generative AI has enabled the creation of novel antibiotics capable of killing two highly drug resistant infections.
Massachusetts Institute of Technology (MIT) scientists described in the journal Cell how they harnessed artificial intelligence to combat a strain of gonorrhoea and MRSA.
It is believed to be the first occasion on which scientists have used AI to design specific antibiotics.
Their approach involved examining nearly 40 million compounds and using AI to identify hypothetically possible molecules or those that have not been confirmed to exist.
Speaking to the Institute’s MIT News, study senior author James Collins said: “We’re excited about the new possibilities that this project opens up for antibiotics development.
"Our work shows the power of AI from a drug design standpoint, and enables us to exploit much larger chemical spaces that were previously inaccessible.”
Collins, professor of medical engineering and science in MIT Institute for Medical Engineering and Science and the department of biological engineering, said the development would be a valuable enhancement in the battle to combat drug resistant ‘superbugs’.
Estimates of the fatalities caused by bacterial infections vary but are thought to run into millions annually.
AI has previously been used to speed the process of identifying the potential of known chemicals for drug use. MIT went a further step by seeking to create new antibiotics and venturing beyond previously identified compounds.
This involved searching some 45 million chemical fragments, initially narrowed down to 4 million using machine learning models to predict antibacterial activity.
Further filters were applied, such as removing anything too closely resembling current antibiotics and those which were toxic to humans.
Continuous analysis identified one particular fragment effective against Neisseria gonorrhoeae. Two generative AI algorithms were then applied to discover additional compounds, producing 7 million overall, subsequently refined to a single compound.
A separate process was used in the case of MRSA but this time using AI to freely design molecules rather than searching chemical fragments.
Production of the antibiotics however will involve lengthy drug development. The designs tested using mice and bacteria require clinical trials prior to application in humans.
Pic: Ahmad Ardity
