Global Interpretability
The ability to understand the overall logic and behaviour of an entire AI model, including which features drive predictions across all inputs.
In plain language
Understanding the big picture of how an AI works overall; what patterns it learned, what it generally considers important. Like understanding a company's overall hiring philosophy rather than just one hiring decision.
Why this matters
Understanding a model's overall behaviour is essential for governance oversight and stakeholder trust. Global interpretability enables risk managers and auditors to assess whether an AI system's logic aligns with organisational policies and ethical principles.
Relevance
GovernanceGlobal interpretability supports governance oversight by enabling auditors and risk teams to verify alignment between AI system behaviour and organisational values.
Related terms
Putting global 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.