The KU Leuven team leading Work Package 5, which focuses on ethical aspects and challenges of AI-based clinical tools, has published a new open-access paper in Frontiers in Public Health entitled “AI-driven tools for the prediction of obesity-related vascular diseases: stakeholder perspectives and challenges”.
The study examines how artificial intelligence could support the early prediction of vascular diseases in people living with obesity. As part of Task 5.2, the team conducted semi-structured interviews with a wide range of stakeholders, including clinicians, patients, researchers, AI developers, and policymakers. Participants were recruited from the AI-POD consortium, professional networks, and relevant European organizations, including obesity associations.
The results show that technical performance alone is not enough for AI tools to succeed in real-world clinical settings. Trust, usability, ethical design and integration into everyday clinical workflows are just as crucial. Stakeholders also highlighted potential risks such as bias, digital exclusion, anxiety, and stigmatisation, emphasising the need to develop AI tools in collaboration with end users and those affected.
The paper underscores the importance of considering ethical and societal aspects early in AI development. By capturing the actual perspectives of stakeholders, the study provides practical guidance for designing AI tools that are not only technically robust but also socially acceptable and clinically relevant.