The opportunities of responsible AI in healthcare
Artificial intelligence (AI) offers significant social and economic opportunities for healthcare. However, these opportunities can only be realized if ethical, legal, social, and economic (ELSA) aspects are systematically anticipated and taken into account in the design and deployment of AI systems. This is particularly urgent in healthcare: AI applications that do not meet societal expectations are often not trusted or accepted by healthcare professionals and patients. This can lead to problems in healthcare, such as misdiagnoses, delayed treatment, and greater health inequalities.
Although the healthcare sector is the largest investment domain for AI worldwide, large-scale implementation in daily healthcare practice is lagging behind. This gap between technological promise and actual application, known as theAI chasm, calls for targeted, value-driven solutions.
The role of AI in medical practice
AI is playing an increasingly important role in diagnostics, decision support, risk prediction, and the organization of care. At the same time, medical AI is confronted with a rapidly changing landscape of ethical guidelines and legal frameworks, such as the EU AI Act and national policy initiatives. Although these frameworks are important for safety and the protection of fundamental rights, they create friction in clinical practice.
It is often unclear whether regulations sufficiently reflect the values and needs of users, how abstract standards should be translated into concrete courses of action, and whether compliance with legal requirements actually leads to the acceptance and use of AI in healthcare. Without attention to these questions, regulations risk hindering rather than facilitating the implementation of valuable medical AI.
About the VAMAI ELSA Lab
The Value Alignment in Medical AI (VAMAI) ELSA Lab focuses on bridging the gap between regulations, ethical ambitions, and everyday healthcare practice. The lab operates on the premise that safe AI alone is not enough: medical AI must actually add value to healthcare and therefore be aligned with human values. VAMAI investigates how the values of patients and healthcare professionals can be identified, analyzed, and embedded in the development, implementation, and evaluation of medical AI. In doing so, the lab combines insights from ethics, law, social sciences, and AI research, and structurally involves patients and healthcare providers in the research process.
The goal of the VAMAI ELSA Lab is to enable medical AI that is not only technically reliable and legally compliant, but also trusted and used in healthcare practice because it demonstrably aligns with the values and needs of patients and healthcare professionals.
Guiding value-aligned medical AI
ELSA Lab VAMAI develops knowledge and practical tools to reduce friction between medical AI, regulations, and clinical practice, thereby enabling meaningful implementation of AI in healthcare. The lab's work includes:
- Understanding which values are central to patients and healthcare professionals in medical AI, and how these relate to personal values, trust, acceptance, and equitable care.
- Analysis of tensions between existing and emerging regulations, such as the EU AI Act, and everyday healthcare practice, and how these tensions can be bridged.
- Development of concrete guidelines and methods to translate abstract ethical and legal requirements into applicable design and implementation choices.
- Application and further development of value alignment and explainable AI in realistic medical use cases, at the individual, organizational, and societal levels.
Collaboration partners
The challenges surrounding value alignment in medical AI call for an interdisciplinary and cross-sector approach. Within the VAMAI ELSA Lab, academic researchers, healthcare professionals, patient representatives, technology developers, and policymakers work together. Through this quadruple helix collaboration, the lab contributes to medical AI that is not only innovative and compliant, but also trusted and used in healthcare practice. There are many collaboration partners, more about which later.
Want to know more or collaborate?
Please contactKarin Jongsma, principal investigator, by sending an email to k.r.jongsma@umcutrecht.nl.
