Using subgroup discovery, an artificial intelligence (AI) approach that identifies statistically exceptional subgroups in a dataset, we develop a strategy for a rational design of catalytic materials. We identify “materials genes” (features of catalyst materials) that correlate with mechanisms that trigger, facilitate, or hinder the activation of carbon dioxide (CO2) towards a chemical conversion. The approach is used to address the conversion of CO2 to fuels and other useful chemicals. The AI model is trained on high-throughput first-principles data for a broad family of oxides. We demonstrate that bending of the gas-phase linear molecule, previously proposed as the indicator of activation, is insufficient to account for the good catalytic performance of experimentally characterized oxide surfaces. Instead, our AI approach identifies the common feature of these surfaces in the binding of a molecular O atom to a surface cation, which results in a strong elongation and therefore weakening of one molecular C-O bond. The same conclusion is obtained by using the bending indicator only when incombination with the Sabatier principle. Based on these findings, we propose a set of new promising oxide-based catalyst materials for CO2 conversion, and a recipe to find more. Our analysis also reveals advantages of local pattern discovery methods such as subgroup discovery over standard global regression approaches in discovering combinations of materials properties that result in a catalytic activation.