FDA Seeks Public Input on Adaptive AI Medical Devices That Can Change Their Answers
OVERVIEWThe FDA is seeking public comment on how to regulate "adaptive AI" medical devices—AI systems that can modify their behavior, outputs, or decision logic based on real-world clinical data after deployment. This regulatory question is central to generative AI medical devices, which may update their models or parameters in response to new information. Current FDA frameworks assume static medical devices with fixed performance characteristics validated pre-market. Adaptive AI devices violate this assumption: they change over time, making pre-market validation incomplete. The FDA's inquiry addresses core regulatory questions: (1) What post-market monitoring is required?; (2) Who is accountable if performance degrades after adaptation?; (3) How should regulators authorize model updates or retraining?; (4) What validation is required before and after adaptation? The FDA's openness to public input reflects genuine uncertainty about how to regulate adaptive AI. This regulatory gap creates deployment ambiguity for vendors and health systems. Organizations should anticipate that FDA will require more rigorous post-market surveillance, transparency about model changes, and potentially preapproval of model updates for high-risk adaptive AI devices. ---
Subscribe to read full wiki analysis
Full wiki content — including source provenance, contradiction tracking, related pages, and analyst notes — is available to Subscriber members.
View Plans & SubscribeNew users get a 7-day trial — no credit card required.