US FDA moves on rules for generative AI medical devices, asks public to weigh in

On August 18, 2026, the US Food and Drug Administration (FDA) published a discussion paper on how to regulate medical devices that use generative AI and opened it for public comment. The move responds to a rapid rise in “GenAI‑enabled” tools, from chatbot-style assistants to image generators, being embedded in diagnostic and treatment‑planning systems.

The paper lays out key questions: how to evaluate and monitor “continuously learning” AI systems whose behavior changes over time; what level of disclosure clinicians and patients should receive about AI involvement; and how to balance premarket review with post‑market safety monitoring. It also raises the issue of how responsibility should be shared between device makers, software providers and healthcare institutions.

The FDA says it aims to build on existing medical device rules while adapting them to cloud‑based, ever‑updating AI software. Because US regulation often becomes a reference point for global health‑tech standards, Japanese medical device makers and hospitals are expected to watch this process closely, as it may shape future expectations for safety, documentation and oversight of AI in clinical use.