In this talk, I will present our lab’s latest research on generative AI models designed to simulate cellular and tissue-level perturbations with unprecedented resolution. These models enable us to ask fundamental questions such as: Which interventions can revert a disease phenotype back to a healthy tissue state? and What perturbations can reprogram cells from state A to state B? By learning causal structure from high-dimensional multi-omics and spatial data, our frameworks can propose actionable interventions, predict patient-specific responses to treatment, and identify the most promising therapeutic targets. I will highlight how these models support target discovery, guide experimental design, and accelerate the development of personalized and precision medicine. Overall, this work demonstrates how generative AI can transform our ability to understand, predict, and engineer complex biological systems.