An ionic decision-maker (IDM), with the remarkable ability to solve multi-armed bandit problems (MBPs), has been developed as a new artificial intelligence technology, which have humanlike decision-making capabilities. Owing to their low durability, the Nafion polymer electrolytes and electrode materials normally found within IDMs do not meet the high-level integration and repetitive operation requirements of such devices. Therefore, in this study, mesoporous SiO2, which is a proton conducting inorganic oxide, has been utilized as the IDM electrolyte. The oxide-based IDMs developed showed good solvability for MBPs and excellent adaptability to dramatic environmental shifts, which were both at similar levels to those found in Nafion-based IDMs. While the adaptation exhibited by the oxide-based IDMs was slightly inferior to that found in Nafion-based IDMs, fluctuations in the correct selection rate (CSR) were significantly improved over those observed in Nafion-based IDMs. Further, the average of reached CSR was enhanced to 0.96 in comparison to 0.92 for Nafion-based IDMs.