From ClingantryAI Health Regulator News

The FDA has finalized a rule classifying cardiovascular machine learning-based notification software as a Class II medical device. Published in the Federal Register on September 11, 2026, the order is effective the same day, and this action places the software under special controls rather than the stricter Class III pathway. The agency states this classification provides reasonable assurance of safety and effectiveness while reducing regulatory burdens to improve patient access. Healthcare organizations using or deploying this type of AI-driven notification tool should review the specific special controls identified in the rule to ensure compliance. This final rule applies to manufacturers and developers of these cardiovascular devices.

On August 18, 2026, the FDA's Digital Health Center of Excellence (within CDRH) issued a discussion paper on how it might regulate generative AI-enabled medical devices, and opened docket FDA-2026-N-7874 on Regulations.gov for feedback through October 19, 2026. FDA says the paper is for discussion only: it is not draft or final guidance and does not set expectations for future marketing submissions.

The paper sketches a two-axis risk framework: one axis is how independently the device acts (from non-directive information, to output that directs an action, to fully autonomous function), and the other is how severe the harm would be if someone relied on a wrong output. FDA says it is considering treating "action-directing" GenAI functions as higher risk than ones that only provide information. For premarket review, it describes a competency assessment modeled loosely on how physicians are trained and evaluated, combining non-clinical benchmarking with clinical confirmation. It also outlines risk-proportionate postmarket monitoring options and raises questions about foundation models and agentic AI systems.

Practical upshot: nothing is binding yet, but if you build or buy GenAI tools that recommend or trigger clinical actions, expect FDA's eventual approach to scale scrutiny with how directive the output is and how much harm a wrong answer could cause, and to expect ongoing performance monitoring after launch.

The FDA has authorized over a thousand AI-enabled medical devices, but its first comprehensive rulebook for how these devices should be designed, validated, and monitored over their whole life — published as a draft in January 2025 — has still not been finalized. Why a hospital should care: once final, it will shape what documentation and ongoing performance monitoring you can demand from device vendors, especially for AI tools whose behavior changes with updates. Reasonable ask of vendors today: whether they're already building to the draft guidance rather than waiting.