Artificial intelligence (AI) has the potential to transform healthcare, improve diagnostic efficiency and take personalised medicine to a new level. However, its full potential in the European Union has yet to be realised due to a fragmented regulatory landscape, uncertainty surrounding the use of patient data and concerns among healthcare professionals about legal liability. These are among the findings of a new in-depth study, “The Use of Data in Public Health: Understanding the Potential for Artificial Intelligence in Medicine”, prepared for the European Parliament’s Committee on Public Health (SANT) by Daniela Spajić, a researcher at the KU Leuven Centre for IT & IP Law.
The development of reliable medical AI directly depends on the availability of large volumes of high-quality health data. Although the European Union has adopted landmark legislation in the form of the European Health Data Space Regulation (EHDS Regulation), which aims to facilitate the secondary use of electronic health data for research and innovation, its practical implementation faces challenges arising from its complex interplay with other EU legal frameworks.
Key Findings of the Study
- Potential versus reality in hospitals: Although AI has demonstrated promising results in applications such as medical imaging analysis (MRI, CT) and disease prediction, the actual degree of its deployment across EU Member States remains difficult to estimate. Healthcare professionals often lack clear guidance on which AI tools are clinically validated and safe, and how they should be integrated into routine clinical practice.
- Complex regulatory landscape: The study highlights legal uncertainties arising from the interplay between the EHDS Regulation, the General Data Protection Regulation (GDPR), the Data Governance Act (DGA) and the Artificial Intelligence Act (AI Act). Key areas of uncertainty include the application of “scientific research” as a legal basis for data processing, the scope of the AI Act’s research exemption, and the practical implementation of data altruism, which enables individuals to voluntarily share their personal data for objectives of general interest.
- Liability burden on healthcare professionals: The current framework for allocating responsibility in healthcare relies on the “standard of care”. However, the use of AI in clinical practice creates legal uncertainty, particularly when AI systems produce erroneous results or healthcare professionals cannot fully assess how an algorithm has reached a particular conclusion. This may expose healthcare professionals to considerable liability risks. To become a valuable tool in clinical practice, AI systems must produce highly reliable and transparent results while ensuring patient safety.
- Lack of national strategies and a catalogue of verified AI solutions: Many EU Member States still lack dedicated national strategies for AI in healthcare and harmonised health-specific AI ethics guidelines. In this context, the study refers to the World Health Organization's (WHO) call for the establishment of a centralised, regularly updated catalogue of verified AI solutions within the EU.
A Call to Action for European Policymakers
The study emphasises that adopting legislation alone is not sufficient. For Europe to become a leader in the responsible adoption of AI in healthcare, cross-regulatory coordination is essential, particularly to clarify and harmonise the interpretation and application of the EHDS Regulation, GDPR and AI Act.
At the same time, policymakers must strengthen the cybersecurity of health data systems, ensure high standards of data quality and interoperability, and provide healthcare professionals with greater legal certainty concerning liability when using approved AI systems in clinical practice