In 1928, Alexander Fleming inadvertently developed the first antibiotic- penicillin. Since then, experts in the field have been creating and shaping new forms of antibiotics, helping save millions of lives worldwide. However, discovering the right combination of compounds to make an effective antibiotic is often a tedious endeavor- a long arduous process of trial and error.
However, in a recent paper from IBM, researchers demonstrate how they have used AI technology to generate designs for theoretical antibiotics that have not yet been created. Already, they have used it two develop two new antimicrobial peptides, and by doing so, exponentially reducing the time it would usually take to complete such a task. The process, in short, begins by amassing all known peptide molecules, and analyzing their properties and patterns. With this data, scientists can "pick and choose" what sort of antibiotic they want to create, and the AI virtually assembles permutations of combinations of peptides, until the desired product is achieved.
You can read the full paper from IBM via Nature Biomedical Engineering to learn more about this monumental leap in bioengineering.