What is the purpose of a bias audit across CPMAI projects?

• A. To increase overall accuracy on the aggregate.
• B. To evaluate fairness across subgroups and document findings.
• C. To reduce dataset size.
• D. To ignore fairness criteria.

Answer: (B)

The key idea here is fairness across groups and transparent reporting. A bias audit in CPMAI projects is done to check whether AI outcomes differ among subgroups defined by sensitive attributes (such as gender, race, or age) and to document what is found. This helps ensure the system isn’t producing discriminatory results, provides evidence for governance and accountability, and guides remediation to reduce disparities. It’s about measuring and reporting fairness, not about boosting overall accuracy, shrinking the data, or ignoring fairness criteria.