• A Collaborative Best Practice Guide for Promoting AI Vendor Transparency in Health Care - The HAIP AI Vendor Disclosure Framework

     Sena Kpodzro, M.P.H., Jee Young Kim, Ph.D., Alifia Hasan, M.B.A., Ciera Thomas, M.P.H., Michael Draugelis, B.S., Scott Morgan Jeffries, M.D., Ashley Beecy, M.D., Corinne Stroum, M.S., Alexandra Valladares, M.S., Ph.D., Zev Eigen, J.D., Ph.D., Danny Tobey, J.D., M.D., David E. Vidal, J.D., Mark A. Lifson, Ph.D., Deirdre Mulligan, J.D., Suresh Balu, M.S., M.B.A., and Mark Sendak, M.D., M.P.P.

    Abstract

    Health care delivery organizations (HDOs) face significant challenges in evaluating vendor-developed artificial intelligence (AI) systems owing to limited guidance on what information to request from vendors during procurement. As a result, vendors often provide inconsistent or incomplete responses, making it difficult to assess clinical relevance, safety, and implementation readiness. To address this gap, we developed the Health AI Partnership (HAIP) AI Vendor Disclosure Framework, a community-informed tool designed to support responsible AI system procurement. The framework identifies essential information across five core domains that HDOs should request � and vendors should disclose � to effectively evaluate vendor-developed AI systems: (1) system capabilities and intended use, (2) system performance and compliance, (3) data stewardship, (4) integration requirements, and (5) life cycle management. By standardizing expectations for the information needed in procurement decision-making, the framework aims to enhance transparency and promote safer health care AI adoption. It serves both as a best practice guide and a customizable resource to support HDOs in procuring vendor-developed AI systems.