Quality & risk management of ethical AI use in human health research

The development of Artificial Intelligence/Machine Learning (AI/ML) has acceler-ated significantly, attracting public attention and investment. While AI promises to revolution-ize healthcare and health research with more accurate diagnostics, personalized treatments, and streamlined processes, it also poses risks to fundamental rights and raises concerns about trustworthiness. Regulatory bodies like the EU are responding with frameworks such as the "AI Act," which establishes different risk levels for AI systems, categorizing medical devices using AI as high-risk but exempting scientific research. Despite increasing use of AI in health re-search, as evidenced by rising approvals for AI-related projects, there is a lack of specific guid-ance on how health research stakeholders should practically assess and monitor AI systems within their ethical obligations. This absence of clear guidance can lead to quality issues in health-AI research.
In this policy brief we explore the complexity of quality and risk assessment of AI system to validate their ethical use. We also propose recommendations for action in order to manage this complexity throughout the research project lifecycle and beyond.

Policy Brief