Counterfactual Fairness
A fairness criterion stating that a decision is fair if it would remain the same in a hypothetical scenario where an individual's protected attribute (such as race or gender) were different.
In plain language
A test that asks: if everything about a person stayed exactly the same except their race or gender, would the AI make the same decision? If yes, it passes the fairness test.
Why this matters
Counterfactual fairness provides a rigorous mathematical framework for testing discrimination in AI systems. Including counterfactual analysis in your AI audit and assurance processes strengthens your ability to detect and remediate hidden bias.
Relevance
GovernanceThis approach offers a concrete method to audit whether protected attributes are inappropriately influencing AI decisions, directly supporting your discrimination risk management.
Related terms
Putting counterfactual fairness 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.