Autoionization of water which gives rise to its pH is one of the key properties of aqueous systems. Surfaces of water and aqueous electrolyte solutions are traditionally viewed as devoid of inorganic ions; however, recent molecular simulations and spectroscopic experiments show the presence of certain ions including hydronium in the topmost layer. This raises the question of what is the pH (defined using proton concentration in the topmost layer) of the surface of neat water. Microscopic simulations and measurements with atomistic resolution show that the water surface is acidic due to a strong propensity of hydronium (but not of hydroxide) for the surface. In contrast, macroscopic experiments, such as zeta potential and titration measurements, indicate a negatively charged water surface interpreted in terms of preferential adsorption of OH(-). Here we review recent simulations and experiments characterizing autoionization at the surface of liquid water and ice crystals in an attempt to present and discuss in detail, if not fully resolve, this controversy.