The comment period on FDA's generative AI device framework closed October 19. The agency's discussion paper, "Considerations for the Regulation of Generative AI-Enabled Medical Devices," sits in docket FDA-2026-N-7874 and came out of CDRH's Digital Health Center of Excellence on August 18.

The paper proposes a two-axis risk assessment framework and a premarket approach built on competency assessment: non-clinical benchmarking of the device, then clinical confirmation before it reaches patients. It also lays out several options for risk-proportionate postmarket monitoring, and it addresses foundation models and agentic systems directly.

None of it binds anyone yet. This is a discussion paper, not draft guidance, and FDA says it is still working through the approaches. For a company with a generative model inside its product, the comments filed by October 19 are the record FDA will draw on when it writes guidance, and the competency assessment language is the part worth reading twice.