The U.S. Food and Drug Administration is developing formal regulatory guidance for medical devices that use generative AI, according to Rick Abramson, director of the FDA’s Digital Health Center of Excellence. Abramson told STAT that the agency plans to provide both broad guidance covering generative AI generally and more specialized guidance for areas where the technology creates particular regulatory complexity.
The move reflects the growing use of generative AI in healthcare, where systems can increasingly assist with clinical documentation, medical decision-making and other healthcare tasks. The FDA is considering a “competency-based approach”, suggesting that regulation could focus more on what an AI system is capable of doing and the risks associated with those capabilities rather than applying identical requirements to every AI-enabled product.
For medical-device developers, clearer guidance could reduce one of the biggest uncertainties surrounding generative AI: how regulators will evaluate systems that can change their outputs dynamically rather than behaving like conventional software. Developers need to understand what evidence, testing, monitoring and controls will be expected when AI models are updated or used in different clinical contexts.
The broader takeaway is that healthcare AI is moving from regulatory experimentation toward a more defined oversight framework. The FDA's eventual guidance could become an important foundation for determining how generative AI is tested, approved and monitored in medical devices. At the same time, the agency's emphasis on both broad and specialized guidance suggests that one universal rulebook may not be sufficient for a technology whose capabilities and risks vary dramatically across healthcare applications.
