FDA AI/ML Software as a Medical Device Framework

The FDA has developed a regulatory framework for AI/ML-based Software as a Medical Device (SaMD). As AI models continuously learn and evolve post-deployment, the framework addresses how manufacturers can update models while maintaining safety and effectiveness. Requires pre-market submission for high-risk devices.

• Pre-submission meeting with FDA for AI/ML-based devices
• 510(k) or De Novo clearance for Class II AI medical devices
• PMA (Pre-Market Approval) for high-risk Class III AI devices
• Predetermined Change Control Plan (PCCP) for AI updates
• Transparency: labeling must disclose AI-based functionality
• Algorithm performance testing across diverse patient populations
• Bias assessment for protected patient characteristics
• Post-market surveillance and performance monitoring

1. 1Determine if your AI software meets the definition of a medical device
2. 2Classify device risk and identify appropriate regulatory pathway (510k/De Novo/PMA)
3. 3Develop a Predetermined Change Control Plan for model updates
4. 4Conduct clinical validation across diverse patient populations
5. 5Engage FDA early via Pre-Submission program