Despite centuries of study, scientists still lack the tools to accurately describe what animals eat in a way that captures the true complexity of their diets. Most existing approaches force species into broad, oversimplified categories such as "carnivore" or "omnivore" that obscure meaningful ecological differences and limit our understanding of how diet has powered the evolution of life on Earth. This project addresses that fundamental gap by developing a new framework for measuring the diets of mammals with unprecedented precision and scale. By combining cutting-edge Artificial Intelligence tools with detailed three-dimensional measurements of mammal teeth, this research will reveal how dietary diversity has evolved across the placental mammal radiation, one of the most spectacular events in vertebrate history. Beyond its scientific contributions, this project advances AI literacy in the biological sciences, produces a large publicly accessible database of mammalian dietary and morphological data, and supports STEM education through partnerships with Florida K-12 teachers and undergraduate internship programs at the University of Florida's Florida Museum of Natural History and George A. Smathers Libraries. This project leverages Large Language Models and advanced Natural Language Processing (NLP) techniques to extract and quantify dietary information from thousands of scientific studies spanning more than two centuries of biological literature. Rather than assigning species to discrete diet categories, the contextual NLP extraction identifies diet as the ranked importance of biologically defined food classes, generating a high-dimensional multivariate dietary dataset for more than 1,000 placental mammal species. A human-in-the-loop validation workflow ensures accuracy and repeatability of AI-assisted dietary rankings. Using a novel multivariate liability threshold model, the project will characterize the tempo and mode of dietary macroevolution and its ecologic