HIV researchers test AI system that optimizes resources across multiple cities at once

A research team built an AI system that helps governments decide where to spend HIV prevention money by modeling how infection spreads between cities, not just within them. The system outperformed traditional single-city models when tested on California and Florida data, suggesting that treating cities as isolated units wastes resources that could be saved by coordinating across regions.

Public health agencies currently optimize interventions city by city, ignoring the fact that people move between jurisdictions and so does disease. This paper shows that an AI system accounting for cross-city movement can reduce infections under the same budget. The practical implication: if health departments adopt this approach, they'll need to coordinate resource decisions across multiple cities instead of running each jurisdiction independently — a shift that requires breaking institutional silos that currently prevent that kind of coordination.