When every public-health system asks the same model

Institutional decisions can become correlated because their errors share upstream dependencies. A highly accurate model can still create systemic vulnerability if many organizations use the same provider, architecture, retrieval stack or evaluation assumptions. Public-health AI governance should therefore monitor concentration and preserve meaningful epistemic diversity. Weavidence Insights. Independent scholarly publication; not externally peer reviewed. AI-assisted drafting and editorial development were used. Pietro Corona remains responsible for the final published claims, source selection, interpretation and approval.