When auditing the transparency of an AI system, which of the following would be the MOST effective way to understand the model's decision-making process?

Question

Options

• AEvaluating the diversity of the training data set
• BAnalyzing the complexity of the algorithms used
• CAssessing the computational cost of the model
• DReviewing the explainability of AI outputs

• A
  11% (3)
• B
  4% (1)
• C
  7% (2)
• D
  78% (21)

Explanation

Explainability refers to the degree to which the internal mechanics or outputs of an AI model can be understood and interpreted by humans. Reviewing the explainability of AI outputs (D) directly answers the question of how and why the model reaches a given decision-using tools like SHAP values, LIME, feature importance scores, or decision rules. This is the most direct and effective method for an auditor assessing transparency. Evaluating training data diversity (A) relates to fairness and bias, not decision logic. Analyzing algorithm complexity (B) may reveal whether a model is interpretable by nature, but does not explain individual decisions. Assessing computational cost (C) is a performance metric unrelated to decision transparency.

No community discussion yet for this question.