Examples
Worked examples
- Is an instance
Reporting that a clinical-risk model satisfies equalised odds across reported race categories within ±2 percentage points
Counter-examples
Looks similar, but isn't
- Not an instance
Stating only that 'the model is unbiased' without specifying the criterion is not a fairness claim in the operational sense
Editorial commentary
Fairness is not a single property but a family of mutually exclusive mathematical definitions; the choice among them is a value judgement that should be disclosed. Reports of AI use in research that bears on people should state which fairness criterion (if any) was evaluated and the results.
References
- Barocas, Hardt, Narayanan 2023 ‘Fairness and Machine Learning’ (textbook)
- Chouldechova 2017 ‘Fair Prediction with Disparate Impact’ Big Data
Also known as
Algorithmic fairness · ML fairness
Machine-readable encodings
Use in your systems
xml
<role vocab="credit"
vocab-identifier="https://casrai.org/dictionary/"
vocab-term="AI fairness"
vocab-term-identifier="https://casrai.org/dictionary/term/ai-fairness" />json
{
"@context": "https://schema.org",
"@type": "DefinedTerm",
"name": "AI fairness",
"identifier": "https://casrai.org/dictionary/term/ai-fairness",
"description": "A property of an AI system whereby its outputs satisfy a defined criterion of equitable treatment across specified groups — common criteria include demographic parity, equalised odds, equal opportunity, and calibration parity — recognising that these criteria are often mutually incompatible.",
"inDefinedTermSet": "https://casrai.org/dictionary/domain/generative-ai-use-and-disclosure/",
"url": "https://casrai.org/dictionary/term/ai-fairness",
"sameAs": [
"Algorithmic fairness",
"ML fairness"
],
"license": "https://creativecommons.org/licenses/by/4.0/"
}
