TLDR: The U.S. Food and Drug Administration (FDA) has issued a Request for Information (RFI) seeking public comments on how to measure and evaluate the real-world performance of AI-enabled medical devices. The agency is particularly interested in strategies to detect, assess, and mitigate performance changes, such as ‘data drift,’ to ensure these devices remain safe and effective throughout their operational lifespan. Comments are due by December 1, 2025.
The U.S. Food and Drug Administration (FDA) has initiated a significant move towards enhancing the oversight of artificial intelligence in healthcare by issuing a formal Request for Information (RFI) on the real-world performance of AI-enabled medical devices. This request, announced on September 30, 2025, and reported widely by October 6, 2025, invites public feedback on critical aspects of these advanced technologies.
The primary objective of the RFI is to gather insights into current, practical approaches for measuring and evaluating the safety, effectiveness, and reliability of AI-enabled medical devices once they are deployed in clinical settings. The FDA emphasizes the need for robust strategies to identify and manage performance degradation over time, a phenomenon often referred to as ‘data drift.’ This drift can potentially lead to issues such as reduced reliability, bias, or a decline in the device’s intended performance.
Currently, many AI-enabled medical devices undergo evaluation primarily through retrospective testing or static benchmarks. While these methods are useful for establishing an initial understanding of a device’s baseline performance, the FDA acknowledges that they are ‘not designed to predict behavior in dynamic, real-world environments.’ The agency highlights that ongoing, systematic performance monitoring is increasingly recognized as crucial for maintaining the safe and effective use of AI by observing how these systems behave during actual clinical deployment.
This initiative builds upon discussions held during the November 2024 meeting of the FDA Digital Health Advisory Committee, where AI stakeholders explored various strategies for evaluating AI devices in clinical practice. The FDA’s request includes specific questions to guide commenters, such as what metrics or performance indicators are currently used to measure safety, effectiveness, and reliability, and how these metrics are defined and weighted across different performance and safety dimensions.
Also Read:
- AI Regulatory Non-Compliance Expected to Drive 30% Surge in Tech Legal Disputes by 2028
- G7 Cyber Expert Group Urges Vigilance on AI’s Dual Impact on Financial Cybersecurity
The FDA’s announcement underscores the dual nature of AI, including generative AI, in medicine. While it presents ‘opportunities to improve patient outcomes, advance public health and accelerate medical innovation,’ it also introduces ‘new considerations… particularly with respect to assessing their performance, safety and reliability after deployment in real-world settings.’ The deadline for submitting comments to the FDA is December 1, 2025, as the agency seeks to build a comprehensive framework for the post-market surveillance and continuous assurance of AI-enabled medical devices.
