AI in Medical Devices: Navigating the Regulation in the US, UK, and EU

The integration of Artificial Intelligence (AI) into medical devices represents one of the most promising yet challenging frontiers in healthcare innovation. As manufacturers increasingly incorporate AI and machine learning capabilities into medical devices, regulatory bodies worldwide are adapting their frameworks to ensure these technologies remain safe and effective while not stifling innovation.

This post examines how three important territories for life science development —the United States, the United Kingdom, and the European Union— are currently approaching the regulation of AI-enabled medical devices.

WATCH: Peter Brady, CEO of UOVO explains why regulation of AI powered medical devices requires special treatment.

The United States: FDA's evolving approach

Since the 1990s, the FDA has been taking steps to address the unique challenges posed by AI in medical devices with their adaptive algorithms and continuous learning systems. Their regulatory framework builds upon the existing medical device classification system while incorporating new considerations specific to AI.

The FDA's traditional regulatory pathways – 510(k), De Novo, and PMA (Pre-Market Approval) – remain the foundation for AI-enabled medical devices. However, the agency has introduced additional considerations specifically for AI/ML-based Software as a Medical Device (SaMD) through various initiatives.

• Predetermined Change Control Plans (PCCPs): The FDA has introduced PCCPs to allow manufacturers to specify planned AI/ML modifications in advance, enabling iterative improvement while maintaining regulatory oversight.
• Total product lifecycle approach: This framework aims to provide appropriate oversight throughout the entire lifecycle of AI/ML-based medical devices.
• Transparency: As of August 2024, the FDA has authorised 950 AI/ML-enabled medical devices and maintains a public list to promote transparency.

European Union: MDR, IVDR, and AI Act framework

The EU's regulatory landscape for AI-enabled medical devices has evolved significantly with the implementation of the Medical Device Regulation (MDR), In Vitro Diagnostic Regulation (IVDR), and the recent introduction of the EU AI Act.

WATCH: Peter Brady, CEO of UOVO explains the EU approach to the regulation of AI enabled medical devices.

Key elements of EU regulation

The MDR and IVDR have established a comprehensive framework for medical devices, including those incorporating AI:

• Specific rules for software classification, with AI-enabled devices often falling into higher risk categories
• Mandatory clinical evaluation for AI components
• Enhanced post-market surveillance requirements
• Specific documentation requirements for algorithm development and validation
• Requirements for transparency in AI decision-making processes

EU AI Act integration

The EU AI Act, which entered into force on August 1, 2024, introduces additional regulatory layers for AI systems, including those used in medical devices. Key aspects include:

• Risk-based classification system for AI systems, with most medical AI applications likely falling into the "high-risk" category.
• Strict requirements for high-risk AI systems, including enhanced transparency and risk management processes
• Mandatory conformity assessments by notified bodies for high-risk AI systems

The AI Act will be fully applicable from August 2, 2026, with an additional year for high-risk AI systems to comply. This creates a dual-regulatory framework that manufacturers must navigate alongside the MDR and IVDR.

Implications for manufacturers

• AI-enabled medical devices classified above Class I (under MDR) or Class A (under IVDR) will be subject to the AI Act's requirements.
• Manufacturers must prepare for coordinated conformity assessments under both the AI Act and MDR/IVDR regimes
• Enhanced documentation and transparency requirements for AI decision-making processes in medical devices
• Increased focus on risk management and post-market surveillance for AI components in medical devices

UK: Post-Brexit Innovation

The United Kingdom's post-Brexit approach to regulating AI in medical devices demonstrates a commitment to fostering innovation while maintaining patient safety. The Medicines and Healthcare products Regulatory Agency (MHRA) has taken significant steps to create a regulatory framework that addresses the unique challenges posed by AI as a medical device (AIaMD).

The UK Approach: Innovation first

The UK's strategic approach to the regulation of AI powered devices currently emphasises:

• A proportionate, risk-based approach to regulation
• Specific guidance for adaptive AI algorithms
• Enhanced requirements for algorithmic transparency and explainability
• Support for innovation through regulatory sandboxes

The MHRA's "Software and AI as a Medical Device Change Programme" demonstrates the UK's commitment to creating a regulatory environment that promotes innovation while ensuring patient safety.

UK: Regulatory innovation with "AI Airlock"

Post-Brexit, the UK has introduced innovative approaches to regulating AI-powered medical devices. In December 2024, the Medicines and Healthcare products Regulatory Agency (MHRA) launched its "AI Airlock" pilot scheme to test new regulatory frameworks in collaboration with technology developers and the NHS.

Key features of the AI Airlock pilot

• Testing innovative technologies: Five novel AI-powered medical devices have been selected for evaluation. These include tools for cancer diagnosis, chronic respiratory disease management, and radiology services.
• Focus on safety and effectiveness: The pilot aims to address challenges in generating real-world evidence for adaptive AI systems while ensuring patient safety throughout a device's lifecycle.
• Regulatory flexibility: Findings from this pilot will inform future UK guidance on AI medical devices and influence how UKCA marking is applied.

Implications for manufacturers

The UK's emphasis on innovation support through regulatory sandboxes like the AI Airlock demonstrates its commitment to fostering cutting-edge technologies while maintaining high safety standards.

Cross territory considerations

Despite differences in specific requirements, several common themes emerge across all three territories:

1. Risk-based classification systems that consider both the intended use and the AI functionality

2. Enhanced requirements for transparency and explainability

3. Focus on continuous monitoring and post-market surveillance

4. Recognition of the need for specific protocols for algorithm updates

Key differences

Important distinctions include:

• The FDA's more developed framework for algorithm modifications
• The EU's stricter requirements for clinical evidence
• The UK's emphasis on regulatory flexibility and innovation support

Practical implications for manufacturers

Documentation requirements

Manufacturers developing AI-enabled medical devices for multiple markets should prepare:

• Comprehensive technical documentation covering algorithm development
• Clinical evidence appropriate for the strictest jurisdiction
• Market-specific risk management documentation
• Procedures for post-market surveillance and updates

Quality Management Systems

Organisations need to adapt their quality management systems to address:

• Algorithm validation and verification processes
• Change management procedures for AI updates
• Data management and privacy requirements
• Incident reporting specific to AI-related issues

Looking forward

The regulation of AI in medical devices continues to evolve rapidly. Several trends are likely to shape future developments:

• Increased harmonisation of international requirements
• Development of specific standards for AI validation
• Enhanced focus on real-world performance monitoring
• Greater emphasis on algorithmic fairness and bias prevention

Conclusion

Successfully navigating the regulatory landscape for AI-enabled medical devices requires a thorough understanding of each territory's requirements and a comprehensive approach to compliance. While the regulatory frameworks continue to evolve, manufacturers can prepare by:

• Implementing robust documentation practices
• Developing comprehensive validation protocols
• Establishing strong post-market surveillance systems
• Maintaining flexibility to adapt to changing requirements

As the field continues to develop, close collaboration between manufacturers, regulators, and healthcare providers will be essential to ensure that innovative AI-enabled medical devices can reach patients while maintaining the highest standards of safety and effectiveness.