Bacteriophages, viruses that infect and kill bacteria, were administered to patients before penicillin existed. A hundred years later, with drug-resistant infections killing more than a million people worldwide each year, not one phage therapy has been approved by the U.S. Food and Drug Administration.
That gap is not due to a lack of trying. Trials are running on several continents, and last year produced the field's clearest randomized signal yet: a phage cocktail that measurably shrank diabetic foot ulcers where bone infection had set in. The obstacles are stranger and more mundane than the biology, which is part of why artificial intelligence has become such an object of interest here.
