The report aligns with ongoing EU efforts, including the AI Act and the European Health Data Space, and offers ten policy recommendations to policymakers aimed at ensuring patient-centered, effective, safe, fair, and transparent AI in healthcare.

Annexed policy actions cover robust evaluation of AI’s added value, responsible deployment guidelines, expanded ethical frameworks, implementation challenges, AI as a learning health system, interoperability and data sharing, training and education, stakeholder engagement and public trust, defining strategic research priorities, and strengthened governance and oversight.

EASAC and FEAM publish a call for responsible AI use in European healthcare, arguing for treating AI as a healthcare measure with proven benefits, understood risks, and ongoing performance monitoring.

Prominent figures cited include Eva Fialová of the Czech Academy of Sciences and Andr é Knottnerus, EASAC Co-Chair, highlighting leadership and expert input on regulatory frameworks.

Two-thirds of Czech hospitals are currently using AI tools, chiefly for diagnostics and imaging evaluation, with rapid uptake across healthcare applications.

Emphasis on transparency, representative data, and avoidance of biased training to safeguard patients and maintain trust in AI-enabled care.

The report notes AI in healthcare is outpacing regulatory oversight, raising risks such as data breaches, algorithmic bias, and potential loss of human control, and calls for proactive governance.

Warning that development outpaces regulation, highlighting risks of data breaches, bias, and loss of human oversight in sensitive healthcare settings.

Recommendations stress integrating AI provisions with existing regulations (Medical Devices Regulation, GDPR, European Health Data Space), ensuring real clinical benefit, high-quality representative data, meaningful patient explanations, and clear liability to avoid dependence on any single provider.

A push for proactive, harmonised EU-wide regulation to prevent fragmented standards and enable safe, scalable AI across member states.

The report is intended to inform the EU AI Act, European Health Data Space, and other forthcoming legislative initiatives, authored by a 22-member European scientific working group.

Policy recommendations include testing AI in regulatory sandboxes, investing in research with carefully selected datasets, and pairing deployment with transparent evaluation and governance.