FDA Floats Competency-Based Testing for Generative AI Medical Devices

The U.S. Food and Drug Administration is asking the public to weigh in on how it should regulate medical devices that use generative artificial intelligence, including whether such products should pass doctor-style competency tests before they reach the market. The agency's Digital Health Center of Excellence, part of the Center for Devices and Radiological Health, published a discussion paper and request for feedback on August 18.

The FDA said generative AI-enabled devices "hold transformative promise for patient care and the broader health ecosystem," but may also introduce risks that differ from those of traditional software and earlier AI-enabled devices.

Four topics, and what the paper is not

The paper poses questions on four topics: how to assess the risk of GenAI-enabled devices, how to evaluate them before they are marketed, how to monitor them once they are in use, and other regulatory issues. Respondents do not have to answer every question and may address only the areas relevant to their expertise, according to the FDA's page.

The agency was explicit about the document's limits. It said the paper "does not represent draft or final guidance," is not intended to propose or implement policy changes, and does not communicate the evidence CDRH expects in future marketing submissions. It also does not address whether the approaches it discusses fall within the FDA's existing legal authorities or would require new ones.

Comments from device makers, clinicians, researchers and the public are due by October 19, 2026, under docket number FDA-2026-N-7874 on Regulations.gov.

Testing the product, not the model

The central idea is a competency-based approach to premarket review. According to Nextgov/FCW, which reviewed the paper, CDRH wrote that it "is considering a competency-based approach to premarket evaluation of GenAI-enabled devices that is also inspired, at a high level, by how human clinicians are evaluated and credentialed, but would be adapted for the technical, practical, and legal considerations applicable to the regulation of medical devices."

The FDA would test the final, user-facing product rather than the foundation model or its subcomponents, Nextgov reported, and the evaluation could use scoring rubrics and expert adjudication. How demanding the test would be would depend on a product's risk profile, which could take into account whether the product steers a user toward an action or only provides information, and what happens if a user relies on a faulty output. The paper also suggests testing alone may not be enough and that the agency could collect real-world evidence after a product is in use, according to the report.

Rick Abramson, director of the Digital Health Center of Excellence, framed the paper as a starting point. "With this paper, we are putting our ideas into the public square and inviting challenge and input from the broader ecosystem so that we can refine these concepts before translating them into policy," he wrote on LinkedIn, as quoted by Nextgov.

The paper follows meetings of an FDA advisory committee on generative AI in 2024 and 2025 and several earlier agency requests for information on AI, Nextgov reported. The American Medical Association is separately weighing whether AI licenses are the right path, according to the outlet.