As the rates of carbon dioxide emissions have continually increased on a global scale, mainly due to high demand of fossil fuel burning for transportation and generating electricity, an inevitable interest in managing and recycling the CO₂ byproducts has emerged. New solutions are focusing on sustainable reduction of CO₂ to CO, which can be conveniently converted to useful hydrocarbon liquid fuel products, such as methane, using existing technologies and infrastructure, as well as renewably generated electricity. Of particular interest in this field is the development and improvement of effective CO₂ reduction catalysts, which ceria-based materials have been shown to be suitable for intermediate temperature solid oxide fuel cells. However, understanding the physical mechanisms by which ceria-based catalysts reduce CO₂ has posed a challenging problem. In this work, recently-developed continuum approaches for the near-surface chemistry of high-temperature mixed conductors and data driven model-building routines have been combined to gain a quantitative understanding of the physical phenomenon of high-temperature CO₂ reduction on ceria catalysts. The main insights pertain to the surface and near-surface defect structure of these materials, where the presence of oxygen vacancies is linked to the enhanced catalytic activity. The experimental data is in-operando scanning surface potential probe measurements performed on specially-designed patterned and thin-film test cells at temperature and under polarization. Preliminary results of data-driven model building will be presented and discussed.