Require targeted testing and a documented human-review: Before using the tool in screening, what requirement
Rare, severe employment harms require targeted testing and a defined human-review path.
The question
An employer is piloting an AI screening tool. A particular applicant error is expected only rarely, but it could exclude qualified people from employment. The vendor reports strong average accuracy. Before using the tool in screening, what requirement is most important?
Preparing for AIGP? Take the free 5-min readiness quiz →
- Track average accuracy throughout the pilot.Average accuracy can conceal the rare exclusion error that creates the serious employment harm described.
- Have recruiters informally watch unusual outcomes and escalate concerns when noticed.Informal observation does not define testing, review authority, evidence requirements, or a reliable response to rare severe errors.
- Compare screening speed with the employer’s manual process and report the savings.Processing speed measures efficiency but does not address whether qualified applicants are wrongly excluded.
- Require targeted testing and a documented human-review path for severe errors. ✓Targeted testing examines the identified failure mode, while a documented review path gives competent people a defined way to investigate or correct harmful screening outcomes.
The trap
For consequential decisions, evaluate the specific severe failure mode rather than relying on average accuracy. How to remember it
Rare, severe employment harms require targeted testing and a defined human-review path.
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
- Prioritize the severe missed-fraud scenario before: Which risk should receive priority? →
- Set risk-acceptance thresholds before the pilot: What should governance require before approving the pilot? →
- Measure correct eligibility decisions: Which evaluation best reflects the affected-person outcome? →
- 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.