The FDA solicited public input on approaches to regulating generative AI-enabled medical devices, focusing on how to assess risk and what evidence should support premarket evaluation and postmarketing monitoring. The agency’s request is designed to clarify how developers should document performance, safety, and changes over time as models evolve. The consultation includes questions about managing uncertainty, defining what constitutes acceptable validation, and tailoring requirements based on clinical risk and intended use. For developers, it signals that generative AI will likely face requirements beyond standard “locked” model evaluation. With adoption accelerating across clinical workflows, the FDA’s approach could influence timelines for new submissions and the design of postmarket surveillance programs tied to real-world performance.