The FDA is seeking input on how to regulate generative AI-enabled medical devices, outlining questions it says will shape future policy on risk assessment, premarket evaluation and postmarketing monitoring. The agency’s request is aimed at clarifying how developers should document model behavior and safety as these systems move from research into clinical workflows. The consultation centers on whether risk categorizations can map cleanly onto models that change or behave differently across inputs, and what evidence regulators should expect for performance claims. It also asks how manufacturers should monitor real-world outcomes once devices are deployed. For biotech and digital health developers building AI-driven diagnostics and decision support, the FDA’s signals are likely to influence development plans, documentation requirements, and clinical validation strategies for the next generation of regulated systems.