Question 1: Single choice
mediumAI-checkedResponsible AI & FairnessGroups A and B are equally represented among credit card applicants. A model with group-specific approval thresholds approves about as many applicants from A as from B, and its accuracy is about the same for both groups. It approves every Group A applicant who should be approved, but only 63% of the Group B applicants who should be approved. Which of these criteria does it violate?