The U.S. Food and Drug Administration (FDA) has released a draft guidance document titled “Artificial Intelligence-Enabled Device Software Functions: Lifecycle Management and Marketing Submission Recommendations.” Published on January 7, the guidance aims to provide manufacturers with comprehensive strategies for designing, developing, and maintaining artificial intelligence-enabled device software functions (AI-DSFs) throughout their total product lifecycle.
This nonbinding draft serves as an essential resource for manufacturers navigating regulatory pathways, emphasizing FDA’s commitment to fostering safe, effective, and high-quality AI medical devices while addressing concerns like transparency, data bias, and ongoing performance monitoring.
The guidance underscores a lifecycle management approach, integrating design, development, deployment, and post-market considerations. Manufacturers are encouraged to employ risk assessments, data management protocols, and performance validation to ensure the safety and efficacy of AI-DSFs across various user demographics.
FDA emphasizes designing devices that promote transparency and minimize bias. Recommendations include collecting evidence to evaluate whether devices perform effectively across diverse demographic groups. For example, training datasets must be representative to prevent systematic inaccuracies in device predictions.
The draft outlines specific documentation and information manufacturers should include in marketing submissions, such as:
- Detailed device descriptions.
- Information on AI model development, including training data sources and performance validation.
- User interface designs and labeling that ensure usability and convey critical safety information.
Acknowledging the dynamic nature of AI, the guidance supports the use of Predetermined Change Control Plans, allowing manufacturers to make certain modifications to AI-DSFs without additional FDA authorization, provided the changes align with pre-approved plans.
For combination products incorporating AI-DSFs, the guidance encourages early engagement with FDA’s lead review divisions to address unique challenges and ensure consistency in regulatory submissions.
FDA is soliciting public comments on the draft for 90 days following its publication in the Federal Register. Stakeholders can submit their input electronically or in writing, helping shape the final version of the guidance.
Need Help?
If you’re concerned or have questions about how to navigate the U.S. or global AI regulatory landscape, don’t hesitate to reach out to BABL AI. Their Audit Experts can offer valuable insight and ensure you’re informed and compliant.
