Efforts to develop sustainable and nutritious food alternatives have recently faced challenges because designing foods that are both nutritious/sustainable and desirable at the same time requires a thorough understanding of the human preferences in food. In addition to common food attributes like flavor and aroma, this also includes mechanical properties of the foods such as texture and rheology [1,2]. In this study, we attempt to characterize a subspace of the human palate, focusing on burgers, using generative artificial intelligence (AI). We train a generative AI model using human-designed recipes collected from the internet, and use the resulting model for the discovery and design of novel burgers that are simultaneously sustainable/nutritious and have desirable taste and texture. We find that the generative model learns the trends in the training data very well, with an impressive ability to extrapolate. The generated burgers were validated using sensory feedback from a survey of 100 people from the general population on attributes such as overall liking and flavor, but also commonly probed texture-related qualities of foods such as softness, hardness, and fibrousness [3]. The burgers receive generally favorable feedback despite being optimized for secondary variables like environmental impact and nutritional value. We conclude that generative AI can be used for accelerating innovation and discovery in food science and the foods industry by enabling a rapid exploration of the human palate, and enabling optimization of various attributes such as mechanical properties, palatability, sustainability, and nutrition.