Turning the EU AI Act into action for health across Europe

Europe’s most ambitious regulatory framework is in full motion. Regardless of shifting enforcement deadlines, implementation readiness in Member States will be the determining factor in whether the EU achieves its ambitious goals. The EU AI Act represents a historic opportunity to make AI in health safer, more trustworthy, and more beneficial for patients across Europe. A new HealthAI report, launched this week on the margins of the World Health Assembly, offers a strategic roadmap for turning that ambition into action.

“The EU has matched ambitious regulatory frameworks with significant policy architecture designed to position itself as a global AI leader. However, our analysis identifies critical implementation gaps.”
— Dr. Ricardo Baptista Leite, CEO of HealthAI

Our new report, Harnessing AI for Health and Economic Competitiveness: Translating the EU AI Act into Action, examines how the European Commission and four of its largest Member States, Germany, France, Italy, and Spain, are navigating the intersection of AI innovation and regulatory compliance. Based on 20 stakeholder consultations and policy analysis conducted between October 2025 and April 2026, it offers a clear look at where implementation stands and what needs to change.

What We Found

Mismatch between EU programs’ timelines and the EU AI Act enforcement: the implementation infrastructure required to deliver on EU ambitions is arriving later than the compliance obligations it is supposed to support, including harmonization of health data through EHDS, harmonized standards, and the Apply AI Strategy’s support measures for healthcare.

Notified body bottleneck: 51 bodies are designated under the Medical Device Regulation, but the AI Act designation requires demonstrating additional AI and data science competencies.

Diverging national approaches: Spain, Germany, France, and Italy are each pursuing distinct institutional models, with different underlying governance architectures and readiness levels.

Untested supervision models: No country has yet validated its dual supervision approach in practice, creating enforcement uncertainty across the board.

Market access for AI in healthcare: Only Germany and France have structured reimbursement pathways that may enable the scaling of AI tools into clinical practice. In Spain and Italy, approved technologies often lack the budgetary and legal basis for adoption, meaning regulatory approval does not translate into real-world use.

“Building on HealthAI’s expertise working with national governments, we focus our analysis on two crucial elements for successful AI governance implementation in health: institutional coordination and enforcement capacity. While the European Union is in the right direction with its policies and ambitions, we point out concrete measures to strengthen implementation readiness.”
— Amanda Leal, AI Governance and Policy Specialist and Lead Author

Our Five Recommendations

• Test dual-supervision models through joint exercises between AI authorities and medical device regulators before high-risk requirements for AI in medical devices take full effect.
• Harmonize reimbursement pathways across the EU so that regulatory approval actually leads to clinical adoption, especially for SMEs.
• Clarify the relationship between the European Health Data Space, the AI Act, and medical device regulation through coordination instruments and roadmaps.
• Create structured peer-learning channels among Member States, leveraging each country’s comparative strengths through DG SANTE and the AI Office.
• Embed multistakeholder engagement in implementation decisions on oversight structures, conformity assessment, and enforcement priorities.

Developed in collaboration with strategic partners, including the World Economic Forum, Friends of Europe, UNITE Parliamentarians Network for Global Health, the European Health Forum Gastein, and with the kind support of Roche and Philips, the report marks a significant milestone in HealthAI’s mission to serve as the global implementation partner for responsible AI governance in health.

The full report is available to download on our Knowledge Hub.