Quiz: Sampling Evidence: Model Cards, Logs, Datasets and Approvals

Quiz Summary

0 of 5 Questions completed

Questions:

You have already completed the quiz before. Hence you can not start it again.

Quiz is loading…

You must sign in or sign up to start the quiz.

You must first complete the following:

Results

Quiz complete. Results are being recorded.

Results

0 of 5 Questions answered correctly

Your time:

You have reached 0 of 0 point(s), (0)

Question 1 of 5

1. Question

What is the primary purpose of examining model cards during an AI management system audit?

Model cards are marketing materials and are not relevant to audits
Correct Incorrect Correct answer

Model cards document what an AI system is, what data it was trained on and its limitations, providing evidence of Annex A.4 requirements
Correct Incorrect Correct answer

Model cards must be identical across all systems
Correct Incorrect Correct answer

Model cards are required only for high-risk systems, not medium or low-risk systems
Correct Incorrect Correct answer

Correct / 1 Points

Incorrect / 1 Points
Question 2 of 5

2. Question

An auditor reviews a model change log and finds that three of the past ten changes lack any approval record. How should this be reported?

The finding should not be reported because some unapproved changes are normal
Correct Incorrect Correct answer

The finding should be reported as a major nonconformity because approvals are required and were not documented
Correct Incorrect Correct answer

The finding should be noted only if more than half the changes were unapproved
Correct Incorrect Correct answer

The finding is irrelevant because the systems still work
Correct Incorrect Correct answer

Correct / 1 Points

Incorrect / 1 Points
Question 3 of 5

3. Question

Why is sampling datasets by requesting an inventory and interviewing data engineers more effective than attempting to read every data row?

Because no audit time can be spent on data
Correct Incorrect Correct answer

Because reading every row is impossible, but asking about provenance, validation and known limitations reveals data governance adequacy
Correct Incorrect Correct answer

Because data governance is not relevant to AI management systems
Correct Incorrect Correct answer

Because datasets do not contain any useful audit evidence
Correct Incorrect Correct answer

Correct / 1 Points

Incorrect / 1 Points
Question 4 of 5

4. Question

An auditor interviews a data scientist about model training procedures. The data scientist explains steps that differ from the documented policy. How should this discrepancy be handled?

Accept the interview explanation and ignore the documented policy
Correct Incorrect Correct answer

Report it as a finding showing actual practice differs from documented procedure
Correct Incorrect Correct answer

Assume the policy is correct and the data scientist is mistaken
Correct Incorrect Correct answer

Do not mention the discrepancy because the model still performs well
Correct Incorrect Correct answer

Correct / 1 Points

Incorrect / 1 Points
Question 5 of 5

5. Question

What should an auditor do if a control requires documented information but the document does not exist?

Assume the control is not applicable to that organisation
Correct Incorrect Correct answer

Record it precisely as a nonconformity citing the clause requirement and documenting that the required information is absent
Correct Incorrect Correct answer

Wait until the next audit to address the gap
Correct Incorrect Correct answer

Ask management to create the documentation retroactively so the finding disappears
Correct Incorrect Correct answer

Correct / 1 Points

Incorrect / 1 Points

Quiz: Sampling Evidence: Model Cards, Logs, Datasets and Approvals

Quiz Summary

Information

Results

Results

Categories

1. Question

2. Question

3. Question

4. Question

5. Question