FDA Signals New Regulatory Approach for AI Medical Devices
A new FDA discussion paper on generative AI in medical devices abandons familiar terminology and proposes a "competency-based" evaluation model, signaling a potential overhaul of its regulatory framework.
The US Food and Drug Administration has released a discussion paper outlining a potential new framework for regulating generative AI-enabled medical devices. The paper is notable for omitting established terminology, such as 'software as a medical device' (SaMD), and introducing novel concepts for evaluating the technology’s unique risks and benefits. Sophisticated counsel should note the proposal for a 'competency-based approach' to evaluation, which the agency likens to the credentialing of human clinicians, focusing on underlying knowledge rather than exhaustive testing of every possible scenario. This represents a potential paradigm shift from current review processes and could fundamentally alter the pathway to market for AI-driven medical technology. The FDA acknowledges that implementing such a framework would likely require statutory changes. Stakeholders have an opportunity to shape the future of this regulation, with the FDA seeking feedback on 26 specific questions until October 19, 2026.