AI engineering glossary · AI ethics law product
What is label bias?
Label bias (also called annotation bias): Distortion in the target labels of a dataset, caused by human annotators' subjective judgements, inconsistent guidelines or prejudiced past decisions. A model treats labels as ground truth, so any bias in them is learned and amplified.
Label bias: an example
Annotators rate the same sentence as more 'toxic' when it is written in a dialect they are unfamiliar with, so a toxicity classifier flags that dialect unfairly.
Related terms
- Measurement bias
- Ground truth
- Algorithmic bias
- Inter-annotator agreement
More AI ethics law product terms
- Lawful basis
- LIME
- Local explanation
- Machine unlearning
- Model card
- North Star metric
- OECD AI Principles
- Overreliance on AI
Where you learn it
Label bias is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.