Local Interpretability
The ability to understand why an AI model made a specific prediction for a particular input instance, as opposed to understanding the model's general behavior.
In plain language
Understanding why the AI made one specific decision. Not how it works in general, but why it said "no" to this particular customer on this particular day given their specific circumstances.
Why this matters
Local interpretability supports individual-level accountability and transparency. When a customer or regulator asks why a specific decision was made, your governance framework needs to provide a clear, documented, case-specific explanation. This is essential for fairness, trust and regulatory compliance.
Relevance
GovernanceLocal interpretability enables transparent accountability by providing case-specific justifications for individual AI decisions.
Putting local interpretability 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.