Machine learning framework predicts pathogen risks in drinking water sources

Beneath every glass of tap water lies an invisible calculation of risk—one that conventional monitoring has long struggled to complete. Researchers in Eastern China have now trained machine learning models to predict the presence of dangerous pathogens in source water using only the routine measurements utilities already collect, bridging a critical gap between what water systems know and what they need to know. The work arrives at a moment when the limits of traditional bacterial indicators—particularly their failure to track viral threats—have left a quiet but consequential blind spot in pub