Counterfactual Situation Testing: Uncovering Discrimination under Fairness given the Difference
This addresses discrimination detection in AI systems for fairness auditing, but it is incremental as it builds on existing legal and causal frameworks.
The paper tackles the problem of detecting discrimination in classifiers by introducing counterfactual situation testing (CST), a causal data mining framework that extends situation testing to answer what the model outcome would be if an individual had a different protected status, and it shows that CST uncovers more cases of discrimination than situation testing, even when classifiers meet counterfactual fairness conditions.
We present counterfactual situation testing (CST), a causal data mining framework for detecting discrimination in classifiers. CST aims to answer in an actionable and meaningful way the intuitive question "what would have been the model outcome had the individual, or complainant, been of a different protected status?" It extends the legally-grounded situation testing of Thanh et al. (2011) by operationalizing the notion of fairness given the difference using counterfactual reasoning. For any complainant, we find and compare similar protected and non-protected instances in the dataset used by the classifier to construct a control and test group, where a difference between the decision outcomes of the two groups implies potential individual discrimination. Unlike situation testing, which builds both groups around the complainant, we build the test group on the complainant's counterfactual generated using causal knowledge. The counterfactual is intended to reflect how the protected attribute when changed affects the seemingly neutral attributes used by the classifier, which is taken for granted in many frameworks for discrimination. Under CST, we compare similar individuals within each group but dissimilar individuals across both groups due to the possible difference between the complainant and its counterfactual. Evaluating our framework on two classification scenarios, we show that it uncovers a greater number of cases than situation testing, even when the classifier satisfies the counterfactual fairness condition of Kusner et al. (2017).