Unit 2: Data Governance, Fairness & Bias Handling
Bias can sneak into AI systems in many ways, even with good intentions:
Historical Bias: Old data reflects real-world inequality
Sampling Bias: Over/underrepresentation of groups
Labeling Bias: Human assumptions in training labels
Measurement Bias: Incomplete or misleading success metrics
Exclusion Bias: Leaving out relevant data or identities
