Document the disparity as a risk: Which action best fulfills the objective of managing issues and risks during
Managing a test-stage fairness risk means documenting it, investigating the cause, and mitigating before release rather than deferring or masking it.
The question
During validation testing of a hiring-screening model, the team discovers the model selects one gender group at a materially lower rate despite comparable qualifications. The system has not yet been released. Which action best fulfills the objective of managing issues and risks during training and testing?
Preparing for AIGP? Take the free 5-min readiness quiz →
- Proceed to deploy on schedule and add the fairness concern to the post-deployment monitoring backlog to be revisited during the next quarterly maintenance review.Plausible because monitoring is a real control, but deferring a known pre-release fairness defect to production shifts avoidable risk onto affected candidates instead of managing it at the test stage.
- Raise the model's overall decision threshold uniformly so fewer candidates pass, reducing the visible count of adverse outcomes across every group at once.Plausible because threshold tuning is a legitimate lever, but a uniform shift masks the disparity rather than diagnosing and correcting its cause, leaving the underlying bias intact.
- Document the disparity as a risk, investigate its cause in the data and model, and apply mitigation before release rather than proceeding to deployment. ✓Correct because managing test-stage risks means recording the issue, diagnosing its source, and remediating it before release, keeping the risk from reaching production.
- Remove the gender attribute from the feature set and redeploy immediately, assuming that dropping the sensitive field is sufficient to eliminate the observed disparity.Plausible because excluding the attribute seems protective, but proxy features can reproduce the disparity, so acting without investigation does not reliably manage the risk.
The trap
Believing that deferring a known defect to monitoring, masking it with thresholds, or dropping the sensitive attribute manages the risk, when proper handling requires investigation and mitigation before release. How to remember it
Managing a test-stage fairness risk means documenting it, investigating the cause, and mitigating before release rather than deferring or masking it.
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
- Bias testing, which examines the model's outputs across: Which test type is specifically intended to detect →
- Record the datasets: Which practice best meets the objective of documenting the training and testing process? →
- A model card summarizing the model's intended use: Which artifact should they produce? →
- 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.
