The U.S. Food and Drug Administration (FDA) has updated its list of AI-enabled medical devices authorized for marketing in the United States, in a step aimed at improving transparency for innovators, health care providers, and patients. The page, published on September 3, 2026, presents the devices in reverse chronological order according to the date of the final decision, while the most recent decision currently shown in the data is dated June 29, 2026.
The list does not itself constitute a new regulatory authorization; rather, it compiles devices that have met the applicable premarket requirements, including focused reviews of safety and effectiveness and the suitability of studies for the intended use and the device’s technical characteristics. Each entry is linked to a submission number and the device’s entry in the FDA database, where publicly releasable information, such as safety and effectiveness summaries, can be found.
A Reference List, Not a Comprehensive Registry
The FDA explains that the list is not a comprehensive resource for all medical devices that use artificial intelligence. The listed devices were identified primarily through the appearance of AI-related terms in descriptions of marketing authorization document summaries or in the device classification, based on the FDA’s Digital Health and Artificial Intelligence Lexicon.
This limitation is important when interpreting the data: a device’s absence from the list does not necessarily prove that it does not use artificial intelligence, and the public summaries linked to database entries do not include most of the information that may have been submitted as part of the authorization application. The page provides options to download the data in CSV, Excel, and XML formats, making it usable for tracking and analysis, while its scope remains tied to the agency’s identification methodology.
Moving Toward Identifying Large Language Models
The FDA says it will explore ways to identify and label devices that incorporate foundation models covering a broad range of AI systems, from large language models (LLMs) to multimodal architectures. The aim is to help innovators, health care providers, and patients determine whether a function based on a large language model is present within a device.
The agency encourages device sponsors to include appropriate information in their public summaries to help it develop more user-friendly ways to identify AI-enabled devices in future updates. However, the page does not yet specify a final standard or a date for adding an independent label for devices that use foundation models.
What Does This Mean for Users and the Industry?
The list provides a practical regulatory signal regarding the U.S. landscape of AI-enabled medical devices and allows comparisons of product names, companies, specialties, product codes, and submission numbers from a single government source. The data show the use of AI across multiple fields, with extensive representation of imaging and radiology devices, alongside products in cardiology, neurology, gastroenterology, dentistry, and other areas.
However, the list’s practical value will remain tied to the limits of disclosure and the inclusion methodology. It does not reveal the complete evidence file submitted to the FDA, nor does it mean that the listed devices are equivalent in function, level of autonomy, or clinical performance. The agency has pledged to update the list periodically and to add devices that received authorization but whose decision summaries were not published during the data-collection period in a subsequent update. The FDA requests that questions and comments be sent to digitalhealth@fda.hhs.gov.