A company deploys a model to recommend job candidates for interviews. They monitor the disparate impact ratio monthly. In the last month, the ratio dropped from 0.9 to 0.7, indicating a potential bias increase. The company wants to investigate. What should they do first?

Retrain the model immediately to fix the bias

Review the model's predictions for the last month to identify any changes in the candidate pool

Ignore the change because the ratio is still above the common threshold of 0.8

Remove the protected attribute from the model

A recruitment firm uses an AI model to rank candidates for job openings. The model was trained on historical hiring data that favored candidates from certain universities. The firm wants to reduce this bias while maintaining the model's predictive accuracy. Which approach should they take?

Use a fairness constraint during model training that penalizes disparate impact

Remove the university attribute from the model

Post-process the model's rankings by manually adjusting scores for candidates from non-prestigious universities

Collect more data from candidates from non-prestigious universities

What is the primary purpose of conducting a demographic and representation analysis on training data?

To identify imbalances in the representation of protected attribute groups

To select the best algorithm for the model

To measure the overall accuracy of the trained model

To determine the optimal number of features for the model

Which activity is part of a discriminatory pattern review?

Selecting the model's programming language

Comparing prediction rates for different demographic groups

Cleaning missing values in the dataset

PMICertified Professional in Managing AI (PMI-CPMAI)

Domain 1Objective 3

Conduct Bias Checks (e.g., Model, Data, Algorithm) PMI-CPMAI Practice Questions (Page 3)

Part of the Support Responsible and Trustworthy AI Efforts domain, which accounts for 15% of the PMI-CPMAI exam. PMI does not publish an official question count, but from its 160-minute exam (~65–105 total, ~10–16 in this domain), expect 2–3 from this objective — we provide 22 practice questions to prepare you well beyond it. (estimate)

22questions here

5free pages

5concepts

15%of the exam

Questions 11–15

11expert · hard
A company deploys a model to recommend job candidates for interviews. They monitor the disparate impact ratio monthly. In the last month, the ratio dropped from 0.9 to 0.7, indicating a potential bias increase. The company wants to investigate. What should they do first?
Select an answer first
12foundation · easy
Which of the following is an example of a protected attribute that would be examined in a demographic representation analysis?
Select an answer first
13application · medium
A recruitment firm uses an AI model to rank candidates for job openings. The model was trained on historical hiring data that favored candidates from certain universities. The firm wants to reduce this bias while maintaining the model's predictive accuracy. Which approach should they take?
Select an answer first
14foundation · easy
What is the primary purpose of conducting a demographic and representation analysis on training data?
Select an answer first
15foundation · easy
Which activity is part of a discriminatory pattern review?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to PMI-CPMAI

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by PMI. “PMI-CPMAI” is a trademark of its owner, used for identification only.