Bias testing to measure error-rate disparities across: Which pairing of test activities most directly
Bias testing addresses group error-rate disparities and interpretability testing addresses decision understandability, matching both stated concerns.
The question
A team is finalizing its training-and-testing plan for a loan-approval model. Legal warns that error rates may differ across protected groups and that reviewers must be able to understand why the model denies an application. Which pairing of test activities most directly addresses these two concerns?
Preparing for AIGP? Take the free 5-min readiness quiz →
- Performance testing to confirm latency under load, and security testing to probe the model's resistance to adversarial inputs and data-poisoning attempts.Performance and security testing are valid plan elements but address speed and attack resistance, not group disparities or decision understandability.
- Bias testing to measure error-rate disparities across groups, and interpretability testing to make individual model decisions understandable to reviewers. ✓Bias testing targets group error-rate disparities and interpretability testing targets understandability of decisions, matching the two stated concerns precisely.
- Unit testing of individual components, and integration testing to confirm the modules work together correctly across the end-to-end training pipeline.Unit and integration tests verify software correctness of components and their assembly, not fairness across groups or explainability of outputs.
- Validation testing on a holdout set, and performance testing to confirm the accuracy metric clears the agreed threshold before the model advances to release.Validation and performance testing confirm aggregate accuracy but can hide group-level disparities and say nothing about interpretability.
The trap
Assuming strong aggregate validation or performance results imply fairness across subgroups and explainability. How to remember it
Bias testing addresses group error-rate disparities and interpretability testing addresses decision understandability, matching both stated concerns.
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
- Documented data lineage and provenance tracing each: Which practice should have been in place to answer the →
- Remove or correct the leaking feature: What is the most appropriate response? →
- Record datasets and versions: Which documentation approach best satisfies all three aims under the →
- 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.
