The U.S. Food and Drug Administration is beginning to rethink how it regulates generative AI in medicine. Rick Abramson, a radiologist by training and director of the agency’s Digital Health Center of Excellence, said these tools are starting to behave less like traditional medical devices and more like clinicians. “We ended up putting out this idea that maybe these devices are starting to behave a little bit more like me as a physician than a heart valve or a defibrillator,” he explained.
That distinction matters because the FDA’s current framework for medical devices was built for hardware and deterministic software. GenAI systems, by contrast, can interpret images, synthesize patient data, and offer recommendations in ways that resemble clinical reasoning. Abramson’s comments suggest the agency is considering whether such tools should be evaluated on the quality of their judgment and decision-making, rather than solely on mechanical performance.
The remarks are preliminary, but they signal a potential regulatory evolution. If the FDA moves toward assessing GenAI like a physician, it would likely require new evidence standards, monitoring practices, and perhaps a different definition of what counts as a safe and effective product. For now, the agency appears to be opening a conversation about how to adapt its tools to a new class of medical software.