Records of the datasets used: Which documentation most directly lets the auditor validate results and confirm
The record of datasets, procedures, metrics, thresholds, outcomes, and issue resolution is what lets an auditor validate training and testing and confirm compliance.
The question
An auditor reviewing a released model asks for evidence that its training and testing were sound. Which documentation most directly lets the auditor validate results and confirm compliance?
Preparing for AIGP? Take the free 5-min readiness quiz →
- The final source code repository on its own, which shows how the system is built but not which tests were run or how the model actually performed.Plausible as engineering evidence, but code alone does not report the tests, metrics, or outcomes.
- The deployment runbook and on-call rotation, which govern how the system is operated in production rather than how it was validated.Plausible as operational documentation, but it addresses running the system, not validating the model.
- A press release summarizing the benchmark scores, which reports the headline numbers without the procedures that produced them.Plausible since it cites results, but unverifiable headline figures do not evidence sound testing.
- Records of the datasets used, the test procedures, the metrics and thresholds, and the outcomes, including how identified issues were resolved. ✓Correct: this training-and-testing record lets an auditor reconstruct what was done and confirm the results and remediation.
The trap
Offering source code, operational runbooks, or headline benchmark figures as evidence that training and testing were sound. How to remember it
The record of datasets, procedures, metrics, thresholds, outcomes, and issue resolution is what lets an auditor validate training and testing and confirm compliance.
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
- Treat the disparity as a material risk and remediate it: Which response best fits risk management during the →
- To document the model's intended use: What is the primary purpose of a model card at release? →
- Real-world data and conditions can drift over time: What is the main reason ongoing monitoring is needed →
- 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.