Home›Biases›IBI-00066
IBI-000665 - Criticalpublie

Discriminatory Proxy Bias

Biais de proxy discriminant

IBI Score
9.0
out of 10
Severity5_critique
Frequencyfrequent
Scope groups5
Detection difficulty5
🇫🇷 Définition (FR)

Variable apparemment neutre servant de substitut statistique à un attribut protégé.

🇬🇧 Definition (EN)

Apparently neutral variable serving as statistical substitute for a protected attribute.

TriAxis Taxonomy

Structural Classification

O
Origin
O2
Computational & Statistical
L
Lifecycle
L3
Measurement
S
Specific
S01
Proxy bias
💡
Concrete Example

Amazon 2018 : système de tri CV pénalise le mot féminin. Corrélation proxy/genre : 0,67.

📊 Detection Metrics

Demographic Parity DifferenceDisparate Impact RatioEqualized Odds Difference

⚙ Alert Thresholds

Disparate Impact Ratio0.8
Equalized Odds Difference0.1
Demographic Parity Difference0.1

📚 Sources & Scientific References

🛡 Impact & Mitigation

⚠ Documented evidence (Studies & Incidents)

  • Acceptation de 3 recommandations IA incorrectes sur 5 en chirurgie robotique : Étude contrôlée à deux phases mesurant le biais d'automatisation chez 19 chirurgiens. Les participants ont accepté 60,6 % des suggestions d'une IA factice incorrecte.
    Impact: Taux d'acceptation des recommandations incorrectes : 60,6% (43/71). Spécificité de rejet : 39,4%. Confiance auto-déclarée : +0,65 (p < 0,001).

🎯 Vulnerable groups impacted

WomenNon-binary personsBlack / Afro-descendant

🛠 Mitigation methodologies

  • SHAP_GROUP : SHAP Group Analysis
  • COUNTERFACT : Counterfactual Fairness
  • DISPARATE_IMPACT : Disparate Impact Test
  • AUDIT_TIERS : Third-party Audit

🏢 Sectors concerned

Finance & Banking (Risk: /5)Housing & Real Estate (Risk: /5)HR & Recruitment (Risk: /5)Justice & Security (Risk: /5)Insurance (Risk: /5)

⚖ Applicable regulations

🇪🇺 Europe
  • •
    AI_ACT_10 : AI Act (UE) 2024/1689
  • •
    AI_ACT_9 : AI Act (UE) 2024/1689
  • •
    RGPD_22 : RGPD (UE) 2016/679
🌍 International
  • •
    ISO_42001 : ISO/IEC 42001:2023
  • •
    UNESCO_2021 : UNESCO - Recommandation sur l'éthique de l'IA
  • •
    OECD_AI_PRINCIPLES : OCDE - Principes IA
  • •
    COE_AI_CONVENTION_INT : Conseil de l'Europe - Convention IA

📋 Metadata

Contributor
Mohamadou SIDIBE
Institution
AI BIAIS INITIATIVE
Version
1
Created on
October 14, 2026
Updated on
November 4, 2026
UUID
239c885a-5b46-462a-ba1b-49aad8064dee

Explore other biases

250 documented biases, classified by the TriAxis taxonomy.