Question $9$ of $40$ You work for a large credit card company. The company has successfully created an artificial neural network that is $98\%$ successful in predicting when someone is going to have trouble paying their bill. When someone is identified as having trouble, the system lowers their limit and blocks large purchases. Yet with millions of customers, there is still that $2\%$ that is misidentified. These customers call in and complain. But since the system was trained on millions of billing statements, no one is sure why the machine saw the patterns. What is the name of this challenge