We develop causal inference methodologies tailored for high-dimensional data (e.g., high-dimensional covariates). Our group leverages generative AI models to learn structured representations that preserve underlying causal relationships. These representations enable 1) Identifiability of causal effects in settings with latent confounding. 2) Counterfactual inference to simulate outcomes under hypothetical interventions. 3) Quantification of uncertainty for causal effect estimates, which is critical for robust decision-making in biomedical and healcare applications.