The AI system is checked against the marks of trustworthy AI, such as being safe, fair, private, secure, and understandable.
Think of it likeTest-driving a car for brakes, steering, and blind spots before trusting it on the highway, not just admiring how it looks in the driveway.
In plain English
Actually test whether your AI tool is safe, fair, private, and understandable before you lean on it, instead of assuming it is fine.