Choosing which fairness measure to prioritize is a moral and ethical judgement rather than a purely technical decision.

Get ready for the GARP Risk and AI Exam with flashcards and multiple choice questions. Each question comes with hints and explanations. Prepare for success!

Choosing which fairness measure to prioritize is about the values you bring to the model, not just the math you apply. In ML fairness, there are multiple formal definitions—each encodes a different idea of what counts as fair. The act of picking one metric over another expresses a normative stance: it reveals which aspects of fairness you deem most important and which trade-offs you’re willing to accept. That makes the decision inherently a value judgement, since it shapes how benefits or burdens are distributed across groups. For example, opting for equal opportunity reflects a priority on equal access to positive outcomes for those who deserve them, even if that means tolerating differences elsewhere. The other options describe concrete technical notions—an explicit fairness criterion, data-bias issues, or problem framing—without capturing the normative choice about which fairness notion should drive the system.