Geographies of Trust: AI, Biomedicine, and the Next Era of Federated and Visiting Data Models

This report examines the rapidly evolving landscape of biomedical research platforms that facilitate secure, privacy-preserving access to sensitive datasets—such as genomic sequences and clinical records—without the need to move, download, or replicate data across multiple locations. These platforms adopt models such as Data Visitation and Federation, which enable computational tools to move to the data—allowing analysis to occur within its original environment or, in the case of centralized platforms that provide analysis tools, within a single, securely governed repository. This paradigm not only enhances data security and compliance but also redefines how researchers collaborate across institutions and borders. The report categorizes these platforms into three primary models—Data Visitation (DV), Federated, and Centralized DV—alongside a Hybrid variant, each offering distinct approaches to data governance, architecture, and scalability. Through comparative tables and decision guides, the report highlights key performance attributes and structural differences, enabling researchers to navigate complex data-sharing environments efficiently. It also examines the transformative potential of emerging AI technologies, including Generative AI, Large Language Models, and autonomous agents, which are poised to redefine how researchers interact with distributed datasets by automating metadata navigation, cohort generation, and the creation of synthetic data. These innovations promise to enhance analysis, streamline governance, and enable federated MLOps pipelines that preserve privacy while scaling collaboration. To meet the growing demands of complex, cross-institutional global health research, AI must be woven into the very architecture of these data visiting systems—enhancing scalability, governance, and analytical capacity while preserving privacy and ethical integrity.