Investigate the disparity's source in data or design: Under III.B.4, what is the most appropriate way to
Investigate the disparity's source, mitigate, retest, and document the outcome before the model advances.
The question
Testing of a hiring model reveals it recommends candidates from one demographic group at a materially higher rate than equally qualified candidates from another. The system is not yet released. Under III.B.4, what is the most appropriate way to manage this issue?
Preparing for AIGP? Take the free 5-min readiness quiz →
- Release the model on schedule and add a disclaimer noting results may vary by group, since informing users of the limitation transfers responsibility appropriately.A disclaimer does not manage the risk; shipping a known discriminatory model despite a warning leaves the harm unaddressed.
- Discard all testing results and restart data collection entirely, since a fairness finding means the dataset is unusable and must be rebuilt from scratch first.Wholesale restart is disproportionate and premature before the disparity's actual source is investigated, wasting effort without diagnosis.
- Investigate the disparity's source in data or design, apply mitigations, retest, and document the issue and outcome before advancing the model. ✓Investigating root cause, mitigating, retesting, and documenting manages the identified fairness risk during testing before release.
- Raise the overall decision threshold uniformly for all groups, since a stricter global cutoff reduces total recommendations and therefore evens out the results.A uniform threshold shift changes volume but not the relative disparity between groups, so it does not resolve the fairness issue.
The trap
Treating a disclaimer or a uniform threshold change as adequate management of a fairness disparity. How to remember it
Investigate the disparity's source, mitigate, retest, and document the outcome before the model advances.
How many of these would you get right?
One of 1581 AIGP questions on Certsqill. Take a free five-minute check and see your score per domain — not one number, but which section to open tonight.
Test your AIGP readiness — freeMore Understanding How to Govern AI Development questions
- Security testing that probes the pipeline and model: Within the training-and-testing plan, which test type →
- A test record listing each test type performed: Applying III.B.5, which documentation practice would let the →
- A model card summarizing intended use: Under III.C.1, which readiness artifact fits this need? →
- All 426 Understanding How to Govern AI Development questions →
Part of the Certsqill AIGP question bank · Understanding How to Govern AI Development ·
Every answer, right and wrong, comes with its own explanation.
