protected attribute
In plain wordsA personal trait, such as race, sex, religion, age or disability, that the law says must not be used to treat people unfairly.
A personal characteristic on which discrimination is prohibited by law, and, in algorithmic fairnessalgorithmic fairness, n.: The effort to make automated decisions fair: to say in numbers what fair treatment means, measure whether a system achieves it, and correct the system if it does not., the variable (often written A) across whose values fairness is measured. Lists vary by jurisdiction and sector: US federal law covers race, color, religion, sex, national origin, age (40 and over), disability and genetic information, among others; the UK Equality Act 2010 names nine protected characteristics; Article 21 of the EU Charter of Fundamental Rights gives an open-ended list. Deleting the attribute from a model's inputs rarely prevents discrimination, because other features act as proxies (see fairness through unawarenessfairness through unawareness, n.: The simple idea of making a system fair by hiding sensitive traits like race or sex from it; it usually fails, because other information gives those traits away.), while measuring or correcting bias usually requires collecting it, which creates tension with data-protection rules on sensitive data.
We hold out the protected attribute from training but keep it in the evaluation set so we can measure disparities.
Origin
The legal idea comes from civil-rights legislation such as the US Civil Rights Act of 1964 ('protected class' became the usual term in later case law). 'Sensitive attribute' and 'protected attribute' became standard machine-learning vocabulary with discrimination-aware data mining from about 2008.
Usage note
The GDPR's 'special categories of personal data' (Art. 9), such as ethnic origin or health, overlap with but differ from protected grounds. The EU AI Act lets providers of high-risk systems process such data under strict safeguards where strictly necessary to detect and correct bias; since July 2026 this rule (formerly Art. 10(5)) is Article 4a, which extends it to deployers and to other AI systems and models.
Related
- algorithmic biasalgorithmic bias, n.: Unfairness in software that judges or serves people: it consistently produces worse results for some groups, such as women, older people or ethnic minorities, often because it learned from skewed or incomplete records.
- disparate impactdisparate impact, n.: When a rule or system that treats everyone alike on paper still lands much harder on one group, such as women or older workers; the law can forbid this even if no one intended it.
- proxy discriminationproxy discrimination, n.: Discrimination that happens when a system cannot see a trait like race directly but relies on something closely linked to it, such as a postcode, and so treats groups unequally anyway.
- fairness through unawarenessfairness through unawareness, n.: The simple idea of making a system fair by hiding sensitive traits like race or sex from it; it usually fails, because other information gives those traits away.
- demographic paritydemographic parity, n.: A fairness test that asks whether a system says yes (approves the loan, shortlists the applicant) at the same rate for every group, whatever other differences exist between them.
- group fairnessgroup fairness, n.: Measuring fairness by comparing how a system treats whole groups, such as women and men, on average: their approval rates, their error rates or what their scores mean.
Sources
- Equality Act 2010 (UK), c. 15, Part 2, Chapter 1.
Defined in law or standards
UK Equality Act 2010, s. 4 ('protected characteristics'); Charter of Fundamental Rights of the European Union, Art. 21
Mentioned in
algorithmic biasalgorithmic bias, n.: Unfairness in software that judges or serves people: it consistently produces worse results for some groups, such as women, older people or ethnic minorities, often because it learned from skewed or incomplete records., algorithmic fairnessalgorithmic fairness, n.: The effort to make automated decisions fair: to say in numbers what fair treatment means, measure whether a system achieves it, and correct the system if it does not., biasbias, n.: A word with several AI meanings: an adjustable number added inside a model; a systematic error in estimates; or unfair treatment of some groups of people by an AI system., counterfactualcounterfactual, n.: A statement about what would have happened if something had been different — 'had I left earlier, I would have caught the train'. AI uses counterfactuals to explain decisions and to measure causes., disparate impactdisparate impact, n.: When a rule or system that treats everyone alike on paper still lands much harder on one group, such as women or older workers; the law can forbid this even if no one intended it., proxy discriminationproxy discrimination, n.: Discrimination that happens when a system cannot see a trait like race directly but relies on something closely linked to it, such as a postcode, and so treats groups unequally anyway., algorithmic recoursealgorithmic recourse, n.: Giving a person who received an unfavorable automated decision a realistic set of actions they could take to get a better result next time, such as paying down a particular debt., biometric categorisationbiometric categorisation, n.: Using AI to sort people into groups — by age, sex, ethnicity and the like — from their faces, voices or other body features. The EU bans using it to guess sensitive traits such as religion or sexuality., concept erasureconcept erasure, n.: Removing a specific idea — such as a person's gender, or an artist's style — from what an AI model can represent or produce, so the model can no longer use or show it., counterfactual fairnesscounterfactual fairness, n.: A fairness test that asks: had this person been of another race or sex, with everything that would have followed from that, would the system have decided the same way?, demographic paritydemographic parity, n.: A fairness test that asks whether a system says yes (approves the loan, shortlists the applicant) at the same rate for every group, whatever other differences exist between them., equal opportunityequal opportunity, n.: A fairness test that asks whether people who truly deserve a yes (who would repay the loan or succeed in the job) get one at the same rate in every group., equalized oddsequalized odds, n.: A fairness test that asks whether a system makes mistakes at the same rates for every group: wrongly flagging innocent people, and wrongly missing real cases, equally often., fairness through unawarenessfairness through unawareness, n.: The simple idea of making a system fair by hiding sensitive traits like race or sex from it; it usually fails, because other information gives those traits away., group fairnessgroup fairness, n.: Measuring fairness by comparing how a system treats whole groups, such as women and men, on average: their approval rates, their error rates or what their scores mean., iterative nullspace projectioniterative nullspace projection, n.: An early method for removing a trait such as gender from a model's internal representations by repeatedly training a simple detector for it and then flattening away whatever the detector relied on., LEACELEACE, abbr.: A one-step mathematical method for scrubbing a concept, such as gender, from an AI model's internal data so that no simple detector can find it, while changing everything else as little as possible., predictive paritypredictive parity, n.: A fairness test that asks whether a system's verdict means the same thing for every group: of the people it flags as high risk, the same share really are, whatever their group.