US FDA and Device AI/ML
The US FDA has announced steps toward a new regulatory framework to promote the
development of medical devices that use advanced artificial intelligence / machine
learning algorithms - AI algorithms that can learn from and act on data. They have
already authorized some devices having AI capabilities. Their AI Good Machine
Learning Practice lists 10 “guiding principles” for ML/AI¹ to apply FDA’s current
authorities in new ways to keep up with the rapid pace of innovation and still ensure
device safety and performance.
The Agency is looking beyond elemental “locked” algorithm AI devices – devices that
don’t continually adapt or learn - to “true” AI - machine learning algorithms that
continually evolve, often called “adaptive” or “continuously learning” algorithms, that
learn through real-world use. The FDA is exploring a framework to allow modifications
to algorithms to be made from real-world learning and adaptation, while still ensuring
safety and effectiveness of the software required for premarket review. They include the
algorithm’s performance, the added concerns for AI / ML software verification and
validation, the manufacturer’s plan for modifications, and the ability of the manufacturer
to manage and control risks of the modifications, including the software’s
"predetermined change control plan".
¹ https://www.fda.gov/medical-devices/software-medical-device-samd/good-machine-
learning-practice-medical-device-development-guiding-principles
-- jel@jelincoln.com