SHAP Values
SHapley Additive exPlanations. A game theory-based approach that assigns each feature an importance value for a particular prediction, providing consistent and locally accurate explanations.
In plain language
A way to show which factors mattered most in an AI's decision. For example, SHAP might show that a loan was denied 40% because of income, 30% because of debt and 30% because of credit history.
Why this matters
From a governance perspective, SHAP values provide auditable evidence of how AI models make decisions. Incorporating SHAP analysis into your model documentation process strengthens your ability to demonstrate transparency and fairness, and supports investigation when outcomes are disputed.
Relevance
GovernanceSHAP analysis should be required as part of model documentation and audit trails, particularly for high-stakes decisions, to provide defensible explanations if decisions are challenged.
Related terms
Putting shap values into practice in your organisation?
Ready to transform your AI strategy?
Partner with Australia's AI strategy and governance specialists. From adoption roadmaps to ISO 42001 audit readiness.