Towards equitable AI for women’s health: accessible data as a catalyst for innovation

Abstract Artificial intelligence (AI) is advancing rapidly, yet its integration into women’s health remains limited by under-representation in clinical datasets, inconsistent data standards, and limited access to multimodal data resources. We map the current horizon of accessible (i.e., open and accessible on request) data that can contribute to AI development for women’s health. Resources include data repositories, cancer registries, biobanks, and published studies. We provide a working definition of ”women’s health”, a table centralising key accessible data sources, and discuss key challenges and opportunities to advance AI research in the field. To support accessibility and reuse, we also provide an open-access online repository of curated datasets. This paper offers a cornerstone for building equitable AI for women’s health—e.g., supporting future assessments of clinically deploya-bility, diversity, pharmacovigilance—and highlights the global AI research in the women’s health ecosystem.