Building Trust in Healthcare AI: The Vision and Initiatives of the Responsible and Ethical AI in Healthcare Lab (REAiHL) in the Netherlands

An article recently pulished on ICT&Health discusses the goals and ambitions of the Responsible and Ethical AI in Healthcare Lab (REAiHL) in the Netherlands, focusing on the need for a medical ethics framework for the adoption of AI applications in healthcare. The Ethics Lab, led by Michel van Genderen (Erasmus MC), Jeroen van den Hoven (TU Delft), and technology supplier SAS, aims to build trust among practitioners and patients in the use of AI models. Jeroen van den Hoven emphasizes that it is important to distinguish two types of trust: first there is confidence in the system's reliable performance in accordance with specifications and secondly there is moral trust in ethical use of a system. Trust in this latter moral sense is a relation between people. 

The article highlights the importance of transparency, explainability, and a protocol landscape for AI in healthcare. REAiHL proposes the establishment of a medical AI registry, analogous to clinical trials, to log and maintain AI algorithms before research or implementation. The Ethics Lab is based on the six core principles outlined by the World Health Organization (WHO) for the use of AI in digital care, and it collaborates with the WHO to ensure responsible AI use in healthcare.

The Ethics Lab involves practical applications, including shadow models, to integrate ethics into workflows and address the challenges of AI adoption. The article mentions three work packages: defining privacy and data considerations, translating theory to clinical practice, and testing the framework in real-world scenarios. SAS sponsors PhD students to contribute to the lab's work, leveraging their global network and knowledge in AI ethics.

The ultimate goal is to prevent situations where non-transparent AI models cause harm to patients and to promote responsible use of data and transparent AI models in healthcare. The article underscores the importance of tangible outcomes and the application of WHO principles, with a vision for widespread adoption of ethical AI practices in the healthcare sector.