Public health experts generally agree that five interventions saved the most lives over the past century: clean water, sanitation, nutrition, vaccines, and antibiotics. Artificial intelligence (AI) has a credible claim to join that list in the 21st century if it is applied to prevention, public health practice, and the social and environmental conditions that shape health. AI is arriving at a critical moment for public health in the U.S.: infectious diseases are emerging and spreading more rapidly, chronic and environmental diseases are increasing, and climate change is amplifying these threats and adding others. The paradox is that the interventions that do the most to reduce the future burden of disease and its cost attract the least investment, because preventing disease produces nothing to sell. Current trends in AI investment, therefore, have the potential to worsen this structural problem.  In this report, we make the case that this paradox can be corrected and that doing so is in the genuine interest of every entity with the capacity to act.