The chemistry of the interface between the metal halide perovskite absorber and the charge transport layer affects the performance and stability of metal halide perovskite solar cells (PSCs). The literature provides several examples of poor PSC conversion efficiency values, when electron transport layers (ETLs), such as SnO2 and TiO2, are processed by atomic layer deposition (ALD) directly on the perovskite absorber. In the present work, we shed light on the chemical modifications occurring at the perovskite surface, during ALD processing of SnO2 and TiO2, in parallel with the evaluation of the PSC cell performance. The ALD processes are carried out on a (Cs,FA)Pb(I,Br)3 perovskite by adopting tetrakis(dimethylamino)tin(IV) and tetrakis(dimethylamino)titanium(IV) as metal precursors and H2O as the coreactant for SnO2 and TiO2, respectively. Perovskite surface modification occurs in the form of an ultrathin PbBr2 layer. Furthermore, in the case of SnO2, halogen molecules are detected at the interface, in parallel with the initial growth of an oxygen-deficient SnO2. Subgap defect states just above the valence band maximum of SnO2 are also detected. These states act as hole traps at the perovskite/SnO2 interface, subsequently promoting charge recombination and deteriorating the performance of the cell. We hypothesize that a redox reaction between the perovskite, or its decomposition products, and the Sn metal center of the ALD precursor takes place: I− and Br− are oxidized to I2 and Br2, respectively, and Sn(IV) is reduced to Sn(II). In contrast, the Ti(IV) metal center does not undergo any redox process, and, as a result, a promising 11% power conversion efficiency is measured with TiO2 as the ETL. This result strongly suggests that TiO2 may be a more suitable ETL, when processed directly on the perovskite absorber.