A new open-access publication, Stakeholder Engagement for Trustworthy AI: Experiences and Results from the RadioVal Project on Breast Cancer Imaging, details the process and outcomes of involving over 150 stakeholders in developing AI-driven prediction tools for breast cancer treatment response.
The study presents a comprehensive, multi-stakeholder engagement strategy grounded in a social innovation approach and guided by the FUTURE-AI principles for trustworthy artificial intelligence. It describes the stakeholder mapping process, three rounds of social innovation sessions, and a dedicated survey and workshop addressing specific FUTURE-AI principles. The findings provide a transferable framework for designing and implementing AI tools that are transparent, reliable, and responsive to real-world healthcare needs.
The manuscript was led by SHINE 2Europe, in collaboration with the Medical University of Gdansk, La Fe Health Research Institute, and the University of Barcelona, and was published as a book chapter in Trustworthy AI in Cancer Imaging Research. The volume, edited by Gianna Tsakou from MAGGIOLI, brings together 13 chapters showcasing best practices for turning theoretical frameworks on AI trustworthiness into practical applications. It offers concrete guidance for designing, validating, and integrating AI systems in cancer imaging, drawing on insights from multiple international research projects.