Create targeted tests for the rare severe condition: Which additional control most directly addresses this
Rare, severe conditions require targeted scenarios because ordinary aggregate testing may not expose them.
The question
A research-administration screening model meets ordinary accuracy, subgroup, privacy, security, and documentation requirements. Historical testing contains few applications involving a rare medical-disclosure pattern that can cause severe denial errors. All routine test prerequisites are satisfied. Which additional control most directly addresses this risk?
Preparing for AIGP? Take the free 5-min readiness quiz →
- Report the model’s average denial accuracy.Average accuracy can obscure rare but severe failures and therefore cannot establish acceptable behavior for the specified condition.
- Create targeted tests for the rare severe condition. ✓Targeted scenario testing deliberately evaluates the uncommon, consequential condition that ordinary representative sampling undercovers.
- Compare processing speed across application categories.Latency comparisons concern efficiency, not whether the model handles the rare medical-disclosure pattern safely and correctly.
- Increase routine test-set size.A larger routine sample may still omit the rare severe condition, leaving the specific high-consequence failure insufficiently evaluated.
The trap
For low-frequency, high-severity cases, look for focused scenario testing rather than larger averages. How to remember it
Rare, severe conditions require targeted scenarios because ordinary aggregate testing may not expose them.
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
- Record model-data version identifiers: What is the most direct missing control? →
- The undocumented lineage prevents reliable provenance: What do these contrasting observations support? →
- Urban-rural performance is comparable in this test set: What does the evidence most directly show? →
- 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.