Calls for risk-based oversight that protects patients while supporting responsible medical device innovation

October 19, 2026 –  – As generative artificial intelligence (genAI) becomes increasingly capable of performing clinical functions, the American Medical Association (AMA) has outlined recommendations to help the U.S. Food and Drug Administration (FDA) modernize how these technologies are evaluated, regulated, and monitored through their use in patient care.

In a letter to Rick Abramson, MD, director of the FDA’s Digital Health Center of Excellence, the AMA responded to the agency’s request for feedback on its Aug. 18 discussion paper on regulating genAI-enabled medical devices. The AMA proposed approaches to evaluating genAI performance in real-world clinical settings, strengthening post market surveillance, and engaging physicians in defining the standard of care.

The AMA’s perspectives emphasized two overarching principles:

  • Regulatory agility cannot come at the expense of safety and effectiveness. While genAI’s evolving capabilities may require novel evaluation methods, regulatory flexibility must preserve the FDA’s responsibility to establish reasonable assurance of safety and effectiveness. Oversight should reflect a device’s intended use and potential for harm and performance throughout its lifecycle, including after deployment.
  • Physician expertise should inform the standards, not serve as the safety net. Physicians should help shape evaluation criteria and clinical performance standards, drawing on their expertise to identify risks and assess clinical relevance. However, clinician oversight cannot substitute rigorous evidence that a device performs safely and reliably in its intended setting. The AMA has offered to work with FDA and other stakeholders to develop best practices for clinician adjudication, ongoing performance review, and identification of clinically consequential AI failures.

“GenAI has the potential to extend the reach of physicians and fundamentally improve how patients receive care. Realizing that potential requires new approaches to evaluating technologies that can perform increasingly consequential clinical functions,” said AMA President Willie Underwood, MD, MSc, MPH. “Physicians have a critical role in defining what good performance looks like, and the AMA is ready to help build the evidence and oversight frameworks that will allow trustworthy AI that improves care. Manufacturers must remain accountable for the performance of their products, and patients and physicians deserve the confidence that these technologies are safe and effective.”

Consistent with its longstanding position, the AMA supports a least-burdensome, risk-based approach to FDA oversight that protects patients while enabling continued innovation. The AMA’s recommendations underscore the importance of meaningful clinical evidence, appropriate safeguards and ongoing performance monitoring as genAI technologies evolve.

The AMA looks forward to continued collaboration with the FDA, technology developers and other stakeholders to advance responsible innovation that improves patient outcomes. For more information on AMA’s work to lead and advance approaches to liability, licensure and regulation as increasingly capable AI assumes clinical functions, please visit the AMA governance-framework-for agentic-AI.

About the AMA
The American Medical Association is the physicians’ powerful ally in patient care. As the only medical association that convenes 190+ state and specialty medical societies and other critical stakeholders, the AMA represents physicians with a unified voice to all key players in healthcare. The AMA leverages its strength by removing the obstacles that interfere with patient care, leading the charge to prevent chronic disease and confront public health crises, and driving the future of medicine to tackle the biggest challenges in healthcare. For more information, visit ama-assn.org.