Key Developments

The European Commission published a comprehensive study on “Artificial Intelligence in healthcare dissects digital health technologies in Europe” on March 17, 2026, marking the latest significant development in EU AI Act implementation. This comes as the Commission simultaneously advances its transparency framework, with a second draft of the Code of Practice on AI-generated content marking and labelling released on March 5, following extensive consultation involving hundreds of stakeholders.

The transparency code, developed through two working groups established in November 2025 and incorporating 187 written submissions, is expected to be finalised by May-June 2026, just ahead of the August 2026 compliance deadline.

Industry Context

The healthcare sector faces particularly acute challenges under the AI Act. Approximately 75% of commercial AI-enabled medical devices currently on the market are radiology-related, with all but one classified as Class IIa or higher under the Medical Device Regulation (MDR). This classification automatically designates most current healthcare AI solutions as high-risk systems under the AI Act.

Implementation readiness remains a significant concern across the EU. At least 12 member states, including major economies like France, Germany, and Ireland, have missed deadlines to appoint competent authorities, while 19 have failed to designate single points of contact.

Practical Implications

Healthcare AI developers and deployers have less than five months to achieve compliance with high-risk AI system requirements. This includes implementing robust risk management systems, ensuring data governance, maintaining human oversight, and establishing comprehensive documentation procedures.

The delayed publication of key guidance documents and technical standards has left organisations with minimal preparation time. Companies operating AI systems in healthcare must prioritise compliance assessments and begin implementation of required safeguards immediately.

For Irish companies, the lack of appointed national authorities creates additional uncertainty about local enforcement and guidance availability.

Open Questions

Critical uncertainties remain around practical implementation guidance for high-risk classifications and specific compliance requirements for healthcare AI systems. The timeline for finalising technical standards and the readiness of national enforcement mechanisms across member states, particularly in Ireland, Germany, and France, continues to pose challenges for industry preparation.

Source: European Commission