Navigating the European AI Act: Key Implications for Medical Device Manufacturers
- 11 hours ago
- 3 min read
The European Union’s Artificial Intelligence Act, adopted in 2024, marks a significant step in regulating AI technologies across member states. For medical device manufacturers, this new regulation introduces important requirements that intersect with existing rules under the Medical Devices Regulation (MDR) and In Vitro Diagnostic Regulation (IVDR). Understanding how these frameworks work together is essential for companies developing AI-enabled medical devices to ensure compliance, maintain market access, and protect patient safety.
What the European AI Act Means for Medical Devices
The AI Act applies horizontally to AI systems across sectors but places particular emphasis on high-risk AI systems. Medical-device artificial intelligence (MDAI) software falls under this high-risk category due to its potential impact on health and safety. This means manufacturers must meet both the MDR/IVDR requirements and the AI Act’s provisions simultaneously.
The AI Act introduces new obligations around:
Risk classification of AI systems
Data governance and quality of training data
Human oversight mechanisms
Transparency and information provision
Post-market monitoring and reporting
Fundamental rights protection
These requirements aim to ensure AI systems in healthcare are safe, reliable, and respect patient rights.
How the AI Act Interacts with MDR and IVDR
The Medical Device Coordination Group (MDCG) issued guidance (MDCG 2025-6) clarifying the overlap between the AI Act and existing medical device regulations. Key points include:
AI-enabled devices must comply with both the MDR/IVDR and the AI Act.
The MDR/IVDR continue to regulate device safety, performance, and clinical evaluation.
The AI Act adds horizontal AI-specific requirements, especially around data and algorithm transparency.
Manufacturers must integrate AI Act compliance into their existing quality management systems.
This dual compliance means manufacturers cannot treat the AI Act as a separate or secondary regulation. Instead, they must build a unified approach that addresses all regulatory demands.
Timeline for Compliance
The AI Act phases in gradually, giving manufacturers time to adapt:
By August 2, 2026, most provisions of the AI Act become applicable.
By August 2, 2027, the classification rules for high-risk AI systems, including MDAI, fully apply.
This timeline allows manufacturers to assess their AI systems, update documentation, and implement necessary controls before full enforcement.
Practical Steps for Medical Device Manufacturers
Manufacturers developing AI-enabled medical devices should take the following actions:
1. Determine Risk Classification
Assess whether the AI software qualifies as a high-risk system under the AI Act.
Confirm classification under MDR/IVDR to understand overlapping requirements.
2. Implement Data Governance Controls
Ensure training, validation, and testing data are high quality, representative, and free from bias.
Document data sources and processes clearly.
3. Establish Human Oversight
Define how healthcare professionals or users can monitor and intervene in AI decisions.
Provide clear instructions and warnings in device documentation.
4. Prepare Technical Documentation
Update technical files to include AI-specific information such as algorithm design, training data, and risk management.
Include transparency measures that explain AI decision-making processes.
5. Conduct Post-Market Monitoring
Set up systems to collect real-world performance data and report incidents related to AI functions.
Use monitoring results to improve AI safety and effectiveness.
6. Update Quality Management Systems
Integrate AI Act requirements into existing quality management frameworks.
Train staff on new compliance obligations and documentation practices.
Examples of AI Act Impact on Medical Devices
Consider a company developing an AI-powered diagnostic imaging tool. Under the MDR, the device must demonstrate clinical safety and performance. With the AI Act, the manufacturer must also:
Prove the AI model was trained on diverse and validated datasets.
Provide clear explanations of how the AI reaches diagnostic conclusions.
Enable radiologists to override AI suggestions if needed.
Monitor AI performance continuously after market release.
Failing to meet these AI-specific requirements could delay market approval or lead to regulatory penalties.
Challenges and Opportunities
The new regulatory landscape presents challenges such as increased documentation workload and the need for multidisciplinary expertise in AI, clinical science, and regulatory affairs. However, it also offers opportunities to build trust with users and patients by demonstrating commitment to safety and transparency.
Manufacturers who proactively align their development processes with the AI Act and MDR/IVDR will gain a competitive edge and reduce risks of non-compliance.
