CPMAI_V7 practice question 4 of 35

In the case that an algorithm you want to use isn't algorithmically explainable, AI systems should try to do the following:

Full set: 100 questions

Answer options

A: Provide a means to have contestability of the algorithm selected

B: Provide a means to interpret AI results so that cause and effect can be represented. Correct answer

C: Provide a means to have a different team on the project

D: Provide a means to reverse-engineer the algorithm to inspect its performance

Correct Answer:

B. Provide a means to interpret AI results so that cause and effect can be represented.

Explanation

When an algorithm is a “black box,” CPMAI recommends supplying post-hoc interpretability—plots, feature importances, counterfactuals—so stakeholders can understand causal factors affecting outputs, satisfying accountability and regulatory demands.

Why Incorrect

A. Contestability alone does not make results understandable.
C. Team changes do not improve algorithm explainability.
D. Reverse-engineering violates IP and may still fail to expose causal logic.

References

1. Cognilytica. CPMAI v7 Methodology Guide, Phase 5, p. 78-79 (“provide interpretability when algorithm is not inherently explainable”).

2. University of California, Berkeley, CS294-129 “Interpretability in ML,” Lecture 1 notes, 2020 (“post-hoc explanation techniques for black-box models”).

3. DARPA XAI Program, “Explainable AI Principles,” 2018, p. 5 (“enable human understanding of cause-effect despite opaque model”).

Keep practicing

Take the free CPMAI_V7 practice test

Ten exam-style questions with answers and explanations, plus the exam facts and study guides.

Start the free practice test