Acidity is a key determinant of chemical reactivity in atmospheric aqueous aerosols and water microdroplets used for catalysis. However, many fundamental questions about these systems have remained elusive, including how their acidity differs from that of bulk solutions, the degree of heterogeneity between their core and surface, and how the acid-base properties are affected by their size. Here, we perform hybrid density functional theory (DFT)-quality neural network-based molecular simulations with explicit nuclear quantum effects and combine them with an analytic model to describe the pH and self-ion concentrations of droplets and films for sizes ranging from nm to μm. We determine how the acidity of water droplets and thin films is controlled by the properties of the air-water interface and by their surface-to-volume ratio. We show that while the pH is uniform in each system, hydronium and hydroxide ions exhibit concentration gradients that span the two outermost molecular layers, enriching the interface with hydronium cations and depleting it with hydroxide anions. Acidity depends strongly on the surface-to-volume ratio for system sizes below a few tens of nanometers, where the core becomes enriched in hydroxide ions and the pH increases as a result of hydronium stabilization at the interface. These results obtained for pure water systems have important implications for our understanding of chemical reactivity in atmospheric aerosols and for catalysis in aqueous microdroplets.