RadioVal project has published an article in the Journal of Medical Internet Research (JMIR) on a participatory approach to defining requirements for trustworthy artificial intelligence (AI) solutions in healthcare.
Led by the SHINE’s team, with contributions from eight other RadioVal partners, the publication is titled “Integrating Participatory Social Innovation Into Requirements Engineering for AI Health Care Solutions: Case Study”. It describes how the project combined stakeholder engagement and social innovation with formal requirements engineering, to ensure that the development and validation of AI solutions considers not only technical aspects, but also clinical, ethical, regulatory and societal needs.
Through an iterative process, RadioVal engaged nearly 150 stakeholders, including healthcare professionals, patients, AI experts, researchers, policymakers, regulators and other relevant actors across several regions of the world. Their input contributed to a final set of 184 actionable requirements, covering aspects such as usability, fairness, traceability, robustness and explainability.
The methodology developed within RadioVal’s Work Package 1 Multi-stakeholder engagement and Social Innovation provides a structured way to translate the perspectives and needs of different stakeholders into concrete requirements for AI development. While developed in the context of AI-based tools supporting breast cancer care, the approach can also inform the development of trustworthy AI solutions in other areas of healthcare.
Read the open access publication here: https://www.jmir.org/2026/1/e87984