The European Union is undergoing a major transformation in health technology regulation — a clear sign of the sector’s vitality.
With the progressive implementation of Regulation (EU) 2024/1689 on Artificial Intelligence (AI Act), medical device manufacturers must now navigate a dual regulatory framework: the MDR/IVDR for medical safety and performance, and the AI Act for algorithmic governance and transparency.
Beyond a simple accumulation of rules, this represents a new paradigm for clinical and regulatory evaluation in Europe.
AI Act and medical devices: aligning dual compliance, not duplicating it
The joint FAQs published in 2025 by the European Commission, the Medical Device Coordination Group (MDCG) and the AI Board confirm a key principle:
any medical device incorporating AI and already covered by the MDR or IVDR automatically falls under the “high-risk AI system” category of the AI Act.
In practice, this means that a diagnostic support software, an automatic detection tool, or any device embedding a learning algorithm must:
- Comply with the clinical and technical requirements of the MDR/IVDR, and
- Demonstrate specific conformity with the AI Act, including data governance, transparency, explainability, robustness, and continuous monitoring.
These obligations are complementary, not competitive. The AI Act acts as a horizontal layer that reinforces clinical compliance processes without replacing existing evaluations.
Concrete impacts on clinical and regulatory strategy
The integration of the AI Act is reshaping how manufacturers design their clinical evaluation strategies. Clinical trials and performance assessments must now go beyond proving safety and efficacy to also demonstrate:
- Algorithmic robustness and result reproducibility,
- Control of bias linked to training data,
- Traceability of AI-driven decisions throughout the device’s lifecycle.
These requirements introduce a new level of complexity: how can one validate a technology that learns, evolves, and adapts?
For authorities and notified bodies, this means adopting a dynamic interpretation of “clinical evidence”, integrating both performance studies and algorithmic validation.
Five key adaptation areas recommended by authorities
According to the latest MDCG guidance (MDCG 2025-6), manufacturers should adapt in five key areas:
- Risk management – Risk management plans must include AI-specific risks (data bias, model drift, training errors).
- Unified technical documentation – Combine MDR/IVDR requirements with new AI Act elements (traceability, transparency, security).
- Post-market surveillance – Implement continuous monitoring and clinical re-evaluation for evolving algorithms.
- Data governance – Justify the quality, representativeness, and GDPR compliance of datasets.
- Human oversight and training – Ensure users understand the system’s limits and decision-making capacity.
These elements redefine the boundaries between clinical evaluation, regulation, and post-market supervision.
BluePharm: your clinical and regulatory partner at the crossroads of AI and medical regulation
At BluePharm, we support manufacturers during this phase of convergence between AI and medical regulation.
Our role is to translate regulatory complexity into a clear, actionable clinical strategy.
We assist companies in:
- Defining clinical evaluation strategies adapted to high-risk AI medical devices,
- Structuring clinical evidence plans that integrate both technical validation and algorithmic performance,
- Anticipating regulatory expectations through a dual reading of MDR/IVDR and the AI Act,
- Adapting post-market clinical surveillance to account for software updates and algorithm drift.
The goal: to ensure the safety and performance of AI-based devices while accelerating their market access in Europe.
Toward convergence rather than regulatory overload
The clarifications recently published by the European Commission mark a step in the right direction. They open the path to harmonization between medical device regulation and algorithmic governance in healthcare.
However, this convergence will only succeed if manufacturers place clinical evaluation at the heart of their AI approach. Assessing high-risk AI systems must go beyond technical validation — it must remain patient-centered, grounded in clinical evidence and guided by ethical medical decision-making.
This is the vision that BluePharm promotes: bridging the gap between data science and clinical science, ensuring that AI-enabled medical devices are safe, effective, and trustworthy.
Useful links
- EU AI Act – Regulation (EU) 2024/1689
- MDCG 2025-6 – FAQs on AI Act / MDR / IVDR
- King & Spalding – Interplay Between EU AI Act and Medical Device Regulations
- Regulation (EU) 2017/745 – Medical Devices (MDR)
- Regulation (EU) 2017/746 – In Vitro Diagnostics (IVDR)
- Article – High-risk medical devices: a crucial challenge at the crossroads of AI and European certification
