Measurement Bias
Bias introduced by the way features or labels are measured, collected or recorded, leading to systematic inaccuracies that affect model performance for certain groups.
In plain language
Errors that creep in through how you gather data. If you only measure customer satisfaction via online surveys, you miss the views of people who prefer phone contact, skewing the AI's understanding of what customers actually want.
Why this matters
Measurement bias can embed unfair outcomes into AI systems without anyone noticing. Your organisation must scrutinise data collection methods during model development and ongoing operation to ensure no group is systematically misrepresented.
Relevance
ImplementationDetecting and correcting measurement bias requires hands-on review of data collection practices at the point of model building.
Related terms
Putting measurement bias 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.