How AI is used in FDA-authorized medical devices: a taxonomy across 1,016 authorizations
A taxonomy of authorized AI devices helps separate clinical purpose, AI function and input data.
What the paper explores
The authors reviewed 1,016 FDA authorizations to classify how AI is used in medical devices. The taxonomy describes differences in clinical tasks and technical functions, including quantitative image analysis and data generation. It provides a more specific vocabulary than treating all medical AI as one category.
How to interpret the evidence
The counts and technology observations describe the study’s historical dataset. They are not a live inventory of FDA authorizations or proof of the performance of any individual device.
Questions for MedTech teams
Describe the product’s actual function, inputs, outputs and role in care before designing its evaluation. Different AI functions call for different evidence questions.
Sources & further reading
Read original paperDOI: 10.1038/s41746-025-01800-1
Bibliographic record and abstract on PubMedThis is a short editorial reading note on external research. The original research is attributed to its authors and does not represent an HMC study, product validation or regulatory determination.
