Unit 9: AI System Design for Trust, Safety & Accountability
Scenario: A rideshare company launches a driver-rating prediction tool.
Problem: The tool assigns lower scores in certain neighbourhoods due to skewed training data.
Fixes:
Introduced bias detection thresholds
Required human review for outlier ratings
Built dashboards for explainability
✅ Result: Improved fairness, restored driver trust, and reduced support escalations.
