Investigate supplier-specific impacts before concluding: What conclusion best fits trustworthy-AI reasoning?
Aggregate accuracy can coexist with subgroup harm; incomplete subgroup evidence warrants investigation before a trustworthiness conclusion.
The question
A manufacturer reports that an inspection model reduces average defect escapes, while one supplier’s parts receive many false alarms. Overall performance covers all suppliers, but subgroup error counts are incomplete. What conclusion best fits trustworthy-AI reasoning?
Preparing for AIGP? Take the free 5-min readiness quiz →
- Withdraw the inspection model from all manufacturing decisions because of the supplier result.The finding warrants investigation, but it does not by itself establish that the entire system is unusable.
- Investigate supplier-specific impacts before concluding that the system is trustworthy. ✓The observed subgroup concern and incomplete error data require context-specific analysis before fairness and harmful-bias risks can be assessed.
- Deploy a separate supplier model immediately, without diagnosing the error pattern.A separate model could be a later mitigation, but the available evidence does not yet establish its necessity or design.
- Treat the supplier concern as immaterial because aggregate accuracy improved.Aggregate improvement does not show that each supplier experiences acceptable outcomes or that subgroup harms are absent.
The trap
Ask whose outcomes the metric includes and whose harms it may conceal. How to remember it
Aggregate accuracy can coexist with subgroup harm; incomplete subgroup evidence warrants investigation before a trustworthiness conclusion.
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 the Foundations of AI Governance questions
- Increase oversight for the redesign’s broader autonomous: What governance implication follows? →
- Recognize sample accuracy while treating reliability as: Which interpretation is most defensible? →
- Use the records for transparency and provide: Which distinction best guides the response? →
- All 337 Understanding the Foundations of AI Governance questions →
Part of the Certsqill AIGP question bank · Understanding the Foundations of AI Governance ·
Every answer, right and wrong, comes with its own explanation.