Summary: Advances in multimodal longitudinal data and artificial intelligence (AI) create new opportunities for cancer etiology research. We envision an AI-powered discovery workflow integrating an interoperable epidemiologic data ecosystem and causal inference frameworks to accelerate the identification of both cancer causes and the converging biological states for prevention.