In my presentation I will explore the critical role of AI in transforming the landscape of health data management across Europe. Drawing from my extensive experience as a researcher in AI and FAIR data infrastructures, as well as my involvement in various international research projects, I will present how AI-driven solutions can revolutionize the curation, integration, and reuse of personal health data. This presentation will highlight the challenges of fragmented health data, the potential of AI to automate data curation processes, and the importance of creating interoperable and patient-centric health records. By showcasing real-world examples and use cases, I aim to demonstrate how these advancements can empower patients and enhance clinical research.

More and more, people start to appreciate the importance and value of data. This is not only accelerated by the ongoing growth and adoption of AI, but also because it is one of the cornerstones of the EU data-strategy. More specifically, the EU data-strategy also encourages to increase the potential of data by supporting data sharing approaches using ‘data spaces’. In his talk Hans will high-light what data spaces are and how they relate to working with health-related data in the EU. He will use a number of examples from his own experience to describe the current state-of-practice and identify challenges and opportunities going forward.

AI offers a lot of potential for health and health demand forecasting. The prerequisites for AI are Findable Accessible Interoperable and Reusable data or in other words data that are FAIR. Data at most public health institutes are not (yet) FAIR. Sciensano started a data strategy and implementation  plan in 2021 from which evolved the creation of a scientific directorate Data Governance in 2024.

This talk will elaborate on the realities of data at public health institutes and its current data governance plans such as an AI cell and charter. Using my previous professional experiences at 2 bureaus of official statistics, I also will elaborate on examples in official statistics that show the power of structured data (the use of registers), their potential for forecasting and the recognition of ‘public statistical’ output as a leverage for accessing data and data quality by complying to the code of practise in official statistics.

Large Language Models (LLMs) have become central to advancing automation and decision-making across various sectors, raising significant ethical questions. This talk is based on a recent research project in which we proposed a comprehensive comparative analysis of the most advanced LLMs to assess their moral profiles. We subjected several state-of-the-art models to a selection of ethical dilemmas and found that all the proprietary ones are mostly utilitarian and all of the open-weights ones align mostly with values-based ethics. Furthermore, when using the Moral Foundations Questionnaire, all models we probed – except for one – displayed a strong liberal bias. Lastly, we causally intervened in one of the studied models and were able to reliably steer the model’s moral compass to showcase different ethical preferences. All of these results suggest that there is an implicit ethical dimension in already deployed LLMs, an aspect that is generally overlooked.

The fact that current LLMs have an implicitly learned moral compass is of particular relevance in critically sensitive contexts, such as in healthcare. I will lay out my forecasts on how the patient-care provider relationship might be altered in the near term and what consequences these findings might have for public health.

The purpose of the presentation will be to give an overview of the regulatory requirements and challenges which apply to manufacturers of AI-based medical devices in the EU. The presentation will start with a discussion of current key considerations for the CE marking of these devices under the EU MDR and IVDR before analysing the impact and challenges posed by the AI Act on future conformity assessments; The presentation will also include some considerations from a healthcare professional perspective on the use of AI-based medical devices in daily practices.

• The EU and US AI Act
• Differences in their implementation for AI in healthcare
• Potential pathways to measure medical and economic impact of AI in healthcare
• Ethical Guidelines in AI-driven healthcare decision-making

Developing AI systems for use in medical settings is a complex, multi-stage process. Relatively few AI developers publish the methodology of such developments. While healthcare organisations and physicians see many approved devices without sufficient evidence, startups and companies developing AI systems are demanding technical, operational and ethical standards. Meanwhile, regulators such as the US Food and Drug Administration (FDA) or the European Medicines Agency (EMA) have approved hundreds of AI-powered systems for use in hospitals and clinics through less rigorous processes than those for drugs.

It has become clear to policymakers, regulators, clinicians, patients and the scientific and technological community that AI standardisation should be implemented to ensure patient rights, fundamental values and the desired impact on healthcare organisations. The presentation aims to share the CHAIMELEON’s findings in this specific topic. A preliminary toolkit with guidelines related to the design, development and validation of AI systems in healthcare will be shared with the audience. Finally, some concrete examples of AI models in the health domain will be presented.

When you realize that we are witnessing a revolution rather than just a new technological advancement, you have an obligation to do everything in your power to integrate it into your organization.

This presentation will outline the roadmap for incorporating Generative Artificial Intelligence at Hospital General de Granollers. It will also include real-world use cases currently being developed and implemented. Attendees will gain insights into the strategic approach taken by the hospital, the challenges faced, and the transformative impact of this cutting-edge technology on healthcare practices.

The discharge process is a critical bottleneck in patient care, often leading to longer hospital stays when aftercare is not arranged in time. In this talk, I will discuss the challenges faced in predicting both discharge dates and necessary follow-up care using AI. I will share insights and solutions encountered throughout my research journey, highlighting practical applications and future directions.

Healthcare professionals worldwide face a significant administrative burden, spending a large portion of their time on documentation rather than patient care. Studies indicate that, on average, over 50% of a clinician’s time is taken up by bureaucratic tasks such as clinical records, discharge notes, and referrals. This not only impacts the quality of care but also contributes to provider burnout, reduced job satisfaction, and compromised patient experience. The inefficiency created by these tasks leads to high levels of stress among healthcare workers and, ultimately, limits the amount of quality time they can dedicate to patients.

In this talk I will present how AI-powered solutions can tackle the documentation burden, allowing clinicians to refocus on patient care. By leveraging natural language processing and voice recognition technologies, AI tools can capture and transcribe consultations, automatically generating accurate medical notes, discharge summaries, and referrals. This technology enables clinicians to produce high-quality, standardized records with minimal effort, saving valuable time and reducing the risk of burnout. Some solutions integrate seamlessly with existing electronic health record systems, enhancing workflow without disrupting familiar routines. Collectively, these AI advancements not only improve efficiency in clinical settings but also enhance patient-provider interactions, allowing healthcare professionals to focus more fully on patient care ultimately improving satisfaction on both sides of the consultation.

The World Health Organization has prioritized artificial intelligence (AI) as a technology that could assist governments meet global commitments to achieve universal health coverage.  Yet AI also raises numerous trans-national, ethical, legal, commercial and social concerns that pose risks to patients, providers, health systems, and societies.  In response, the WHO convened an expert group on the ethics and governance of artificial intelligence in health in 2019. Over the last five years, the Expert Group has issued international guidance that identifies the risks and challenges for the use of AI, ethical principles that should underpin the design, development and use of AI, and models of governance to assure its appropriate use.  This includes additional guidance on the use of generative AI (large multi-modal models in health care), the use of AI for pharmaceutical development and delivery, and the relationship of AI to ageism.  Currently, WHO is examining how research ethics can provide effective oversight of AI-related health research, while also providing support to Member States, providers, and developers to integrate the principles and recommendations from WHO into their on-going use of AI.