Reversible protonic solid oxide cell (P-SOC) operating at intermediate-temperature exhibits excellent potential as a power generation and green hydrogen production device in fuel cell and electrolysis cell modes because of the high conversion efficiency. However, the lack of efficient air electrodes is the main challenge to obtain P-SOC with remarkable performance. Typically, air electrodes should possess high proton, oxygen ion and electron conductivity, outstanding catalytic ability for oxygen reduction reaction and H2O splitting, and also long-term durability. Recently, high entropy oxides (HEO) have become popular due to their various potential applications in terms of outstanding properties, including catalysis ability, conductivity, thermal stability, etc. HEO air electrodes have been confirmed to show good electrochemical performance in P-SOC, but the complex compositions and structure make it difficult to study HEO by traditional experimental methods. Machine learning (ML) has been regarded as a powerful tool in materials research and can solve the drawbacks in the discovery of HEO in a traditional way. In this perspective, we not only discuss the current utilization of HEO in P-SOC but also provide a possible process to use ML to guide the development of HEO.
