The EURAID Framework:

How can hospitals develop and implement artificial intelligence (AI) in a way that truly improves care instead of adding burden to healthcare professionals? A multidisciplinary team of experts led by Prof. Stephen Gilbert from Else Kröner Fresenius Center (EKFZ) for Digital Health at the Faculty of Medicine of TUD Dresden University of Technology (MF-TUD) and University Hospital Carl Gustav Carus Dresden together with Dr. Anke Diehl from University Hospital Essen, has developed a practical framework: EURAID – European Responsible AI development. EURAID offers hospitals structured guidance on the collaborative, human-centered “in-house” development, validation, and implementation of AI systems. The framework defines mechanisms for involving all stakeholders, outlines processes for iterative design, testing, and clinical evaluation, and provides strategies for training and consensus building. EURAID is based on shared goals, European values such as respect for human dignity, protection of fundamental rights and equality, and European laws.

Guidance for AI development and implementation in hospitals

EURAID was developed by stakeholders representing the entire spectrum of healthcare, including medical and psychological professionals as well as experts in digital transformation of hospitals, medical informatics, and quality management. In addition, perspectives from patient and employee representatives, politicians and health insurers, regulators and ethicists, as well as knowledge from research, were incorporated. The framework provides clear, actionable recommendations for hospitals that want to develop and implement AI in a human-centric, responsible, and safe manner – aligned with the EU’s evolving regulatory landscape and shared European values. The latter refer to patient and employee safety, protection of fundamental rights, transparency of AI systems, human oversight, fairness, and non-discrimination. The recommendations were based on input and expertise from German university hospitals in Dresden, Hamburg, Aachen, and Essen; the German Federal Ministry of Health (BMG); the Federal Institute for Occupational Safety and Health (BAuA); health and accident insurance companies (AOK and DGUV); labor unions (ver.di); a law firm; and several research institutions (including the Fraunhofer Heinrich Hertz Institute and the Karlsruhe Institute of Technology, KIT). From TUD and Dresden University Hospital several professors and researchers contributed their expertise during the process, spanning multiple disciplines from medicine, clinical AI and medical informatics to psychology and regulatory science.

AI technologies are accelerating but implementation in hospitals is still slow

Despite rapid advances in AI technologies and recent efforts at EU-level to simplify the rules governing “in-house” developed AI systems – including the European Commission’s proposal of December 16, 2025, to revise the Medical Device Regulation (MDR) and In Vitro Diagnostic Medical Devices Regulation (IVDR) – the implementation of AI in clinical practice remains slow. Many systems fail to integrate into routine care because they are developed and tested without sufficient consideration of location-specific hospital workflows and the needs of healthcare professionals. As a result, the systems are perceived as an additional burden rather than a relief.

How can hospitals develop and implement human-centric AI systems that ensure safety while fostering innovation?

To achieve this balance, EURAID was developed as a practical, expert-driven guide for human-centric digital transformation in hospital settings. EURAID shows how AI systems can be developed collaboratively and with consistent consideration of non-technical criteria. The framework aims to bridge the gap between research and real-world implementation by demonstrating how AI systems can be co-developed and deployed. The approach is illustrated using a system developed “in-house”, as this enables a tailored application for the hospital using the hospital’s own data. By systematically involving all relevant stakeholders throughout the entire AI life cycle, clarifying respective roles and the degree and scope of participation, presenting real-world use cases, and methods for achieving iterative consensus, EURAID offers a unique and practical approach to reconciling different perspectives and regulatory requirements.

AI systems must be developed with healthcare professionals at core

The authors emphasize that successful digital transformation requires a shift from primarily technology-driven to collaborative-organizational thinking. Successful AI implementation in hospitals can only succeed if the people who use these systems on a daily basis and are affected by them are involved from the start. EURAID shows how this can be achieved even in highly complex, time-pressured environments. Applying the framework clarifies stakeholder roles, identifies barriers early and offers approaches for collaborative fixes. It keeps healthcare professionals at its core, not only as system users, but as co-designers, implementers, and monitors, ensuring that AI implementation in healthcare truly benefits the people.

Digital transformation in hospitals can only succeed if it is designed in collaboration with the people who work with the systems on a daily basis. EURAID demonstrates in a practical way how the introduction of AI can be implemented responsibly, safely, and effectively – even in complex clinical structures”, explains Dr. Anke Diehl, Chief Transformation Officer at University Medicine Essen and Head of the Digital Transformation Unit, who provided valuable practical insights from her own digital transformation initiative together with colleagues from Aachen.

To translate EURAID into action, hospitals should conduct internal readiness assessments, establish cross-functional AI governance structures and define role-specific responsibilities covering ethical, legal, technical and clinical oversight.

Digital systems and AI can make a meaningful contribution to reducing the workload of healthcare professionals and improving patient care. At Dresden University Hospital, we are also working on responsibly establishing new digital solutions and gradually integrating them into clinical practice. EURAID provides important, structured guidance for this,” says Prof. Dr. Uwe Platzbecker, CEO of the University Hospital Carl Gustav Carus Dresden.

EURAID exemplifies our faculty’s commitment to developing innovative research in an interdisciplinary manner and ensuring rapid translation into practice. This is how solutions emerge that are medically meaningful, technically feasible, regulatory sound, and societally acceptable,” says Prof. Dr. Dr. Esther Troost, Dean of the MF-TUD.

About EURAID – European Responsible AI Development

EURAID is a pragmatic, solution-oriented framework, compatible with European values (according to Art. 2, EU Treaty, 2012) and regulations. It ensures that barriers to “in-house” AI development and implementation in hospitals are identified early and resolved through collaborative problem solving. The framework was developed through four multi-stakeholder workshops, and in-dept expert consultation. The authors combine a broad spectrum of complementary expertise, including medical device regulation, hospital digitalization management, quality and clinical risk management, medical informatics, occupational work health and safety, and clinical expertise from diverse healthcare professions. This is complemented by expertise in psychology and human-centered AI design, as well as legal, ethical, policy and academic perspectives, together with insights from insurance providers and labor unions.

List of experts from TUD and Dresden University Hospital Carl Gustav Carus, who contributed to EURAID:

• Anett Schönfelder (first author)
• Maria Eberlein-Gonska (retired)
• Jakob N. Kather
• Rebecca Mathias
• Hannah Sophie Muti
• Prof. Hannes Schlieter
• Martin Sedlmayr
• Magdalena Wekenborg
• Stephen Gilbert (senior co-author)