IBI-002123_significativepublie
Gender Amplification Bias
Biais d'amplification de genre
IBI Score
6.5
out of 10
Severity3_significative
Frequencysystematique
Scope groups3
Detection difficulty3
🇫🇷 Définition (FR)
Associations genre-rôle amplifiées par le modèle au-delà de leur prévalence réelle.
🇬🇧 Definition (EN)
Gender-role associations amplified by the model beyond their real prevalence.
TriAxis Taxonomy
Structural Classification
O
Origin
O2
Computational & Statistical
L
Lifecycle
L2
Representation
S
Specific
S18
Group attribution bias
💡
Concrete Example
MS-COCO : 33% des activités culinaires associées à des femmes → 68% dans les prédictions.
📊 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
- •Men Also Like Shopping: Reducing Gender Bias Amplification using Corpus-level ConstraintsJieyu Zhao, Tianlu Wang, Mark Yatskar, Vicente Ordonez, Kai-Wei Chang (2017)
🛡 Impact & Mitigation
🎯 Vulnerable groups impacted
WomenNon-binary personsBlack / Afro-descendantYouth (18-25)Seniors (60+)Motor disabilityAsian personsArab / MENA personsRural areasLGBTQ+ personsWorking classMinority language speakers
🛠 Mitigation methodologies
- AUDIT_TIERS : Third-party Audit
🏢 Sectors concerned
Education & Training (Risk: /5)Media & Social Platforms (Risk: /5)HR & Recruitment (Risk: /5)Justice & Security (Risk: /5)Healthcare & Medicine (Risk: /5)
⚖ Applicable regulations
🇪🇺 Europe
- •AI_ACT_10 : AI Act (UE) 2024/1689
🇺🇸 United States
- •US_TX_TRAIGA : Texas Responsible AI Governance Act (HB 149)
🌍 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
🇫🇷 France
- •FR_CNIL_AI : CNIL - Recommandations IA
🌐 Canada
- •CA_AIDA : Canada - Loi sur l'IA et les données (AIDA)
📋 Metadata
Contributor
Mohamadou SIDIBE
Institution
AI BIAIS INITIATIVE
Version
1
Created on
October 14, 2026
Updated on
November 4, 2026
UUID
d2fdc649-3825-4a3c-99dd-00c24a337807
Explore other biases
250 documented biases, classified by the TriAxis taxonomy.