If the algorithm uses information about protected classes to make automated decisions, Acme must ensure that the algorithm does not have a disparate impact on protected classes in the output. The Fair Credit Reporting Act (FCRA) protects consumers from unfair, inaccurate, and discriminatory treatment by creditors and other businesses that use credit reports. The FCRA prohibits creditors from using information about protected classes, such as race, color, religion, national origin, sex, marital status, age, or because they receive income from a public assistance program, to make decisions about credit. In the case of Acme Student Loan Company, the algorithm is using information about protected classes to make automated decisions about whether to send payment reminder calls. This could have a disparate impact on protected classes, such as people of color or people with low incomes. For example, people of color may be more likely to be identified as being at risk of default, even if they are just as likely to repay their loans as people of other races. Acme Student Loan Company must ensure that the algorithm does not have a disparate impact on protected classes. This could be done by using a variety of methods, such as:
Testing the algorithm for accuracy, fairness, and bias before and after deployment Providing consumers with notice and consent options for the use of their data Allowing consumers to access, correct, or delete their data Implementing accountability and oversight mechanisms for the algorithm Ensuring compliance with applicable laws and regulations