AI in Healthcare

Specific Requirements for Software as Medical Device Using AI

If a (digital) medical device uses artificial intelligence (AI) in performance- or safety-relevant parts, additional AI-specific requirements apply to the Quality Management System (QMS) and Technical Documentation (TD). Although the MDR does not explicitly mention AI, such requirements can already be derived from the current state of the art. In the near future, explicit regulatory requirements will be introduced through the EU AI Act. This course outlines current QMS and TD requirements, presents the expectations of Notified Bodies, and looks ahead to what the EU AI Act will bring and how to prepare.

Course Content

4 Course Modules

Self-paced digital modules with a final Q&A session led by industry experts.

Requirements for Medical Devices Using AI
This module explains how AI is defined, when AI usage triggers additional requirements, what these are, and where they originate—from Notified Bodies’ expectations and the upcoming EU AI Act.

Requirements in the Early Development Phase
Covers requirements to consider early in development, such as human oversight of AI, setting clinically relevant performance parameters, proof planning, and AI-specific risk management.

The Path to Conformity Assessment
Describes how QMS and TD must be expanded to include AI-specific content: data management, AI training, and validation activities. These are then reviewed by a Notified Body.

Deployment and Maintenance
AI-specific requirements don’t stop after product launch. Post-market surveillance and ongoing maintenance—especially for continuously learning AI—are also regulated. This includes additional duties for manufacturers who also operate the devices.