Which statement best reflects best practice for explainability in AI?

• A. Explainability is optional.
• B. Explainability should be ignored if accuracy is high.
• C. Explainability should be used only for regulatory reasons.
• D. Explainability helps to understand and justify decisions.

Answer: (D)

Explainability in AI serves as a bridge between a model’s outputs and the people who rely on them. It provides the reasoning behind a decision, which is essential for trust, governance, and accountability. Even when a model is highly accurate, knowing why it made a particular choice helps users evaluate whether the result aligns with goals, detect potential biases, and identify errors or edge cases. Explanations also support auditing, regulatory compliance, and the ongoing improvement of the system by revealing which factors influenced the decision and how changes to inputs might affect outcomes. In short, explainability adds transparency and justification that are valuable beyond raw accuracy. Treating it as optional, limiting it to regulatory needs, or dismissing it when accuracy is high can leave decisions opaque, erode trust, and hinder safe and responsible deployment.