Can the Machine Learning Model Tell You How It Arrived at a Specific Prediction?

Darpa has been working on explainable ai, called x a i, for a neural nets. And that is like an explainable model. Is a method or technique which is able to peer into the black box and understand exactly how that model works. If we can't get a specific explanation as to why one particular piece of data resulted in one particular outcome from that machine learning model, cald at least interpret it. So knowing how computers are more sensitive to that data will help us make these systems more reliable and trustworthy.