Evidence that the system would disproportionately flag: Which finding from that assessment would most
A finding of disproportionate harm to a vulnerable group is the impact-assessment result that should trigger a redesign.
The question
A public agency is designing an AI system to prioritize benefit-fraud investigations. Midway through design, a reviewer requires an impact assessment. Which finding from that assessment would most appropriately trigger a redesign?
Preparing for AIGP? Take the free 5-min readiness quiz →
- Evidence that inference costs exceed the budgeted cloud spend, which pushes the team to negotiate a larger compute allowance rather than alter the design.Plausible as a project concern, but a cost overrun is a budgeting issue, not an impact-assessment harm.
- Evidence that the model's training run took longer than planned, delaying the internal timeline for launching the initial pilot.Plausible as a schedule risk, but training duration is not a finding about harm to people.
- Evidence that the user interface uses a color scheme some staff dislike, prompting a broad review of the front-end design choices.Plausible as feedback, but interface aesthetics are unrelated to the harms an impact assessment weighs.
- Evidence that the system would disproportionately flag a vulnerable group, creating harm that the current safeguards do not adequately address. ✓Correct: an impact assessment that reveals disproportionate harm to a vulnerable group is exactly the finding that warrants redesign.
The trap
Treating budget, schedule, or cosmetic findings as impact-assessment triggers instead of harms to affected people. How to remember it
A finding of disproportionate harm to a vulnerable group is the impact-assessment result that should trigger a redesign.
How many of these would you get right?
One of 499 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
- To proactively probe the system for weaknesses and failure: What is the primary goal of red teaming in this →
- Documented data lineage and provenance tracing each: Which practice should have been in place to answer the →
- Continuously monitor performance and drift metrics: Which approach best satisfies both the ongoing-monitoring →
- All 135 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.
