Position: Beyond Sensitive Attributes, ML Fairness Should Quantify Structural Injustice via Social Determinants

1. 1Identify social determinants relevant to your ML model using literature reviews and expert consultations
2. 2Collect and preprocess data on these social determinants
3. 3Develop and apply fairness metrics that account for structural injustice
4. 4Evaluate and compare the performance of your ML model using these new metrics
5. 5Refine your model to mitigate structural injustice and improve fairness

Data scientists and ML engineers working on fairness and bias mitigation can benefit from this approach to better understand and address structural injustice in their models, and product managers can use this insight to develop more equitable products

💡 Structural injustice in ML models can be quantified and addressed by considering social determinants, not just sensitive attributes

🚨 ML fairness should go beyond sensitive attributes! Quantify structural injustice via social determinants 📊