Event summary · Brussels · 30 September 2026
Launch of the joint EASAC–FEAM report on AI in European healthcare
Two panels, one question: how do we move AI in healthcare from promise to practice, safely, ethically and for everyone?
The launch event of the joint EASAC–FEAM report AI in European Healthcare: Policy Recommendations for Optimising Added Value and Safe, Ethical, and Inclusive Adoption opened with Birgitta Henriques-Normark (EASAC Vice-President) and Ferry Breedveld (FEAM President) welcoming the audience.
Prof. André Knottnerus (EASAC Co-Chair) and Prof. Luis Martí-Bonmatí (FEAM Co-Chair), who moderated the event, then gave a few introductory words. They recalled how the project came about and highlighted the range of expertise within the working group that made the report possible.
Louise Abboud (FEAM Executive Director) presented the key findings of the report. Following President von der Leyen’s optimistic vision of AI, the report set out to provide cues and safety nets for the development of AI in healthcare. That vision is now being put to the test across Europe, so the report could not have been published at a better time. It highlights the need to pair technical accuracy, which is not deemed sufficient for patient benefit, with clinical value, judgement and insight.
The key findings can be summarised in three points:
Anticipate long-term effects and embed responsibility before deployment.
Prioritise equitable access, transparency, participation and sovereignty.
Harmonise standards, enable interoperable cross-border data use and strengthen collaborative oversight.
The first section concluded with the ten recommendations of the report:
Moderator: Prof. Luis Martí-Bonmatí, Working Group Co-Chair (FEAM)
Panellists: Anca Scortariu (European Commission, DG CONNECT) · Eva Fialová (working group member) · Andreia M. Oliveira (European Commission, DG SANTE) · Miikka Korja (working group member)
Prof. Martí-Bonmatí opened by asking the panellists: ‘Is it the time for this report?’ Anca Scortariu’s response summed up the consensus of the conversation:
“The report is timely, much needed and important for the European Union.”
Anca Scortariu, European Commission, DG CONNECT
Andreia M. Oliveira added that she was confident in the direction taken, while stressing that the existing and promising motivation to move forward does not equate to a job done. As a member of the audience underlined, AI is in constant evolution and requires constant recalibration to keep pace with its most recent developments.
This connected directly to the question put to Miikka Korja. Asked how he foresaw the adoption of AI in local hospitals, he responded that adoption is taking place at a very different pace depending on countries and capacities. He underlined, however, that AI would become part of everyday activities if used to its fullest capacity within the safety nets established in the report. Prof. Martí-Bonmatí rounded off the discussion on patients’ data by adding that ‘data needs to be exchanged between providers and hospitals, on monitoring internal biases, that are often incorporated for a fast versioning of the software’. Miikka Korja added that family members need to ‘become stakeholders to assess the work of AI, as the latter is in continuous evolution’.
Eva Fialová took the question further, noting that “ethical principles tend to be even more conservative than law”, which requires a “close relationship between lawmakers and physicians to facilitate the integration of AI in hospitals”. She pointed to the need for guidelines, templates and rules to safely exchange the data collected.
Patient consent was also addressed as a crucial factor in the validation and accountability chain. Panellists agreed that it should follow existing rules and remain based on informed consent whenever patients’ personal data are used.
These subjects spurred interest in the audience. Prof. Ferry Breedveld, FEAM President, pointed out that guidelines and rules are often subject to personal and individual revision by scientific councils, raising the question: “is there a role remaining for those guidelines as AI approaches us?” Miikka Korja offered a personal view, stating that guidelines are mostly beneficial for those who are less familiar with the subject matter.
The panel concluded with Prof. Martí-Bonmatí asking: ‘what are we missing?’ Reports provide a good overview of a subject as it stands today, while also bringing to light the points that should be taken further in future work. Several were raised during the discussion, such as the societal impact of AI and the balance between trust and distrust in AI.
Moderator: Prof. André Knottnerus, Working Group Co-Chair (EASAC)
Panellists: Sascha Marschang (HOPE – European Hospital and Healthcare Federation) · Bianca Pop (European Patients’ Forum) · David Ríos Insua (working group member) · Miikka Korja (working group member) · Prof. Luis Martí-Bonmatí (Working Group Co-Chair, FEAM)
The discussion opened with Sascha Marschang underlining the timeliness and importance of the report, which directly addresses the core concerns of HOPE’s team regarding the implementation of AI in hospitals: ethics, gradual implementation, and evidence-based and human-centric procedures. Bianca Pop of the European Patients’ Forum (EPF) reported that, in a survey, 98% of patients were fine with the use of AI in the healthcare sector, which is encouraging for the adoption of the report’s recommendations. She also stressed that patients must be involved in the implementation of AI, as they are the end users of the technology. The EPF advocates a patient-centric view that also protects the patient–doctor relationship.
Turning to differences in infrastructure between hospitals, Sascha Marschang indicated that EU collaboration is the key factor in data exchange between infrastructures.
“The closer we can bring the EU guidelines to the operational reality of healthcare facilities throughout the EU, the better.”
Sascha Marschang, HOPE
He added: “through best practices and shared networks can smaller hospitals catch up”.
Prof. Knottnerus raised underrepresentation and selection bias in databases as problems that need solving. David Ríos Insua answered that systems can be trained to identify, resolve and mitigate such defects. He also emphasised the importance of clinical oversight in identifying and recognising potential issues.
The audience was then invited to put questions to the panellists. Responding to a question on IT infrastructure and implementation, Sascha Marschang pointed out that many hospitals invest no more than 5% of their budget in IT, most of it spent on maintaining existing systems. This, he acknowledged, is the real obstacle to implementing AI. He suggested that hospitals should digitalise first before moving on to AI infrastructure: a time-consuming process, but a necessary one to ensure equity within the EU.
Prof. Martí-Bonmatí reinforced these arguments, stating that “one reason for hospitals to be incorporated into the maturity process of healthcare and digitalisation is because AI is changing what hospitals are doing”. Anca Scortariu insisted that the networks some hospitals belong to are key to spreading this idea, and a good opportunity for hospitals to learn from each other, whether on skills, organisation or sustainable solutions.
Prof. Gheorghe Ioan Mihalaş (Academy of Medical Sciences of Romania) stressed, in line with the report, the educational needs of clinicians, medical students and hospital staff to ensure appropriate management and knowledge of AI in healthcare. Other members of the audience highlighted the importance of following up on the report: AI is developing fast, and today’s reality will not be tomorrow’s.
The report, its presentation and the panel discussions were warmly received, and the event closed to a round of applause. FEAM and EASAC will now bring the report’s recommendations to EU and national decision-makers and, as several participants requested, will consider how best to follow up on this work as AI continues to evolve. Both organisations thank the speakers, moderators, working group members and participants.
