FDA Says Draft Guidance on Generative AI Medical Devices Is Coming Within Months

The Food and Drug Administration expects to publish draft guidance on generative AI-enabled medical devices within months, Rick Abramson, director of the agency’s Digital Health Center of Excellence, said at AdvaMed’s The MedTech Conference in Boston, MedTech Dive reported on October 21. The comment period for the FDA’s August discussion paper on regulating such devices closed on October 19, and by the next day the agency had received more than 500 submissions.

What the FDA said

“This is not just an academic paper. It’s not just a thought exercise. We intend to draft policy, and it’ll come soon,” Abramson said, describing the work as a priority at the level of the Health and Human Services secretary and the White House. He did not commit to a date because staff still have to work through the comments. On the device center’s fiscal 2027 priority list, a draft on generative AI-enabled devices for mental disorders sits on the “A list,” while a broader umbrella guidance on risk assessment and evidence generation is under construction. Finalizing the AI lifecycle management guidance drafted in early 2025, aimed at more static AI devices, is another priority.

What the discussion paper proposed

The August 18 paper, which is not guidance and creates no requirements, outlined a possible two-axis framework for assessing risk, a premarket approach built on competency assessment inspired by how physicians are trained and evaluated, and risk-proportionate postmarket monitoring, along with questions on foundation models and agentic AI. The American Hospital Association, in comments filed October 19, backed a flexible, risk-based framework and asked the agency to clarify that generative AI clinical decision support and administrative tools are not subject to FDA oversight.

Why it matters

The mental health draft grows out of advisory committee meetings on mental health chatbots, a topic that has also moved into courtrooms and enforcement, as in the eight-state lawsuit alleging ChatGPT is unsafe for kids and the FTC’s subpoenas over consumer harms. Device makers will also be watching how the guidance fits wider federal AI policy, including the Super Intelligence Task Force’s risk review. Abroad, Britain’s MHRA plans draft guidance on a predetermined change control framework by the end of the year.

What to watch

Whether the mental health chatbot draft arrives first, how heavily the umbrella guidance relies on postmarket monitoring rather than one-off premarket tests, and whether the FDA accepts hospitals’ request to keep clinical decision support and administrative tools outside its oversight.