← AI security in plain English
MEASURE 1
NIST AI RMF / Measure
The team picks the right tests and measurements for the AI system's risks and then actually runs them, instead of guessing.
Think of it likeUsing a real thermometer and a timer to cook the turkey, rather than poking it and hoping, so you actually know when it is done.
In plain English
Choose sensible ways to test and measure how your AI tool is doing, then use them, so you are working from facts instead of gut feeling.