In Module 11 we discussed how miserably wrong things can go, at times, with AI. In this module, we try to sort out how to get some light at the end of the tunnel by digging into what is influencing the output of a machine-learning model and how.