The National Science Foundation awarded a $299,569 EAGER (EArly-concept Grants for Exploratory Research) Project Grant to the University of Rochester to develop an approach that leverages large language models and artificial intelligence to accelerate the discovery of earth-abundant, active, and selective catalysts for the reverse water-gas shift reaction. The project aims to demonstrate that language-based representations can be universally applied to materials discovery processes expressed through text, with the broader goal of educating and exciting students about the use of language models and AI in materials science. The research focuses initially on understanding and developing trimetallic catalysts, which are more difficult to characterize than bimetallic catalysts, making them a good fit for this language-based predictive approach. The project will also assess the effects of experimental artifacts and irreproducible results on the model's performance and integrate the language-based workflow with existing computational methods to extract mechanistic information from experimental procedures.