世界经济论坛-为健康领域的人工智能赢得信任：一条合作之路（英）

Earning Trust for AI in Health: A Collaborative Path ForwardW H I T E P A P E RJ U N E 2 0 2 5In collaboration with Boston Consulting GroupImages: Getty ImagesContentsForeword 3Executive summary 4Introduction 51 Empowering trustworthy AI in health: The urgent need for collaboration 71.1 Global divergences challenge the scaling of AI in health 81.2 The private sector is key to driving progress and standardization 81.3 AI regulations must be crafted to keep pace with innovation 92 The need for a pragmatic approach: Guidelines, sandboxes and 10 post-market surveillance 2.1 Legislation can build a strong baseline for governing AI in health 102.2 Sandboxes provide a safe space in which the private 11 sector can innovate 2.3 Post-market surveillance can help cope with 12 the evolving nature of AI3 The importance of public–private partnerships for AI in health 133.1 The role of public–private partnerships in regulating 13 medical devices, including software 3.2 Private-sector capabilities can help test and 13 operationalize the regulatory process 3.3 Quality assurance resources: An approach to PPPs 15 for independent testing and training Conclusion 16Appendix: A selection of regulatory sandbox inititatives 17Contributors 18Acknowledgements 18Endnotes 19Disclaimer This document is published by the World Economic Forum as a contribution to a project, insight area or interaction. The findings, interpretations and conclusions expressed herein are a result of a collaborative process facilitated and endorsed by the World Economic Forum but whose results do not necessarily represent the views of the World Economic Forum, nor the entirety of its Members, Partners or other stakeholders.© 2025 World Economic Forum. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, including photocopying and recording, or by any information storage and retrieval system.Earning Trust for AI in Health: A Collaborative Path Forward 2Artificial intelligence (AI) holds great promise to transform healthcare – enhancing diagnostics, optimizing workflows and improving health outcomes for all. However, realizing AI’s benefits responsibly demands a fundamental evolution of how health systems, with their diverse set of stakeholders, develop and build trust in innovation. Existing evaluation frameworks – built for products that remain typically unchanged after approval, such as pharmaceuticals and medical devices – are not fully equipped to manage the dynamic, evolving nature of AI technologies. The probabilistic behaviour of certain AI systems introduces new dimensions of uncertainty that traditional, deterministic approaches cannot fully address.To manage these challenges effectively, regulatory models must evolve. Dynamic governance mechanisms such as regulatory sandboxes, life-cycle evaluation and post-market monitoring will be essential to ensure that AI systems remain safe, effective and equitable through