Which principle BEST guides an auditor assessing whether an AI system's outputs are sufficiently explainable for the use case?
Select an answer to reveal the explanation.
Short Explanation and Infographic
Here's the deal — a is correct because explainability requirements should be proportionate to the decision's stakes — a low-risk content recommendation requires less explainability than a model determining eligibility for housing — and the need of affected parties (individuals, regulators) to understand outcomes. B is technically unrealistic for modern ML models.
Full explanation below image
Full Explanation
A is correct because explainability requirements should be proportionate to the decision's stakes — a low-risk content recommendation requires less explainability than a model determining eligibility for housing — and the need of affected parties (individuals, regulators) to understand outcomes. B is technically unrealistic for modern ML models. C narrows explainability to one legal basis. D is an overstatement — post-hoc explainability techniques exist for neural networks.